• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of aemPermissionsJournals.ASM.orgJournalAEM ArticleJournal InfoAuthorsReviewers
Appl Environ Microbiol. May 2004; 70(5): 2806–2815.
PMCID: PMC404432

New Recombination Methods for Sinorhizobium meliloti Genetics

Abstract

The availability of bacterial genome sequences has created a need for improved methods for sequence-based functional analysis to facilitate moving from annotated DNA sequence to genetic materials for analyzing the roles that postulated genes play in bacterial phenotypes. A powerful cloning method that uses lambda integrase recombination to clone and manipulate DNA sequences has been adapted for use with the gram-negative α-proteobacterium Sinorhizobium meliloti in two ways that increase the utility of the system. Adding plasmid oriT sequences to a set of vehicles allows the plasmids to be transferred to S. meliloti by conjugation and also allows cloned genes to be recombined from one plasmid to another in vivo by a pentaparental mating protocol, saving considerable time and expense. In addition, vehicles that contain yeast Flp recombinase target recombination sequences allow the construction of deletion mutations where the end points of the deletions are located at the ends of the cloned genes. Several deletions were constructed in a cluster of 60 genes on the symbiotic plasmid (pSymA) of S. meliloti, predicted to code for a denitrification pathway. The mutations do not affect the ability of the bacteria to form nitrogen-fixing nodules on Medicago sativa (alfalfa) roots.

Sinorhizobium meliloti is a gram-negative bacterium that is best known for forming nitrogen-fixing symbiotic relationships with legumes, such as alfalfa (25). When not in symbiosis, these rhizobia are part of the normal, free-living soil microflora. S. meliloti cells detect root exudates in the soil and migrate toward, attach to, and invade the root hairs of alfalfa. This process ultimately leads to the formation of specialized root organs called nodules. During the development of nodules, S. meliloti cells differentiate into endosymbiotic forms called bacteroids, which can reduce N2 to NH4+ through an energetically expensive process requiring eight low-potential electrons and at least 16 ATP molecules per molecule of N2 reduced. The bacteroids provide this fixed nitrogen (NH4+) to the plant, and the plant delivers organic acids and perhaps other carbon and energy sources to the bacteroids (6, 10).

The DNA sequence of the S. meliloti strain 1021 genome, which consists of a chromosome (3.65 Mb) and two megaplasmids, pSymA (1.35 Mb) and pSymB (1.68 Mb), has been determined (9). This 6.7-Mb genome was predicted to contain at least 6,207 protein-coding genes along with various insertion sequence elements and phage sequences (9). Having the sequence available is a major step toward describing the biology of the bacteria, but new methods for analyzing the genome are needed in order to obtain a more functional description of the roles played by each of these genetic elements. Manipulating a genome requires extensive use of oligonucleotides, over 12,000 for the S. meliloti genome. The expense of these and the attendant cost of cloning and other standard manipulations are a considerable barrier to working with the entire sequence. However, methods by which genes or other interesting sequences could be cloned initially with a single set of primers and then used in various ways for further studies could be very popular. We sought methods that would be cost-effective, efficient, and accurate and would be adaptable to high-throughput experimental designs.

The methods presented here were developed by modifying and combining existing genetic technologies. We focused on cloning predicted open reading frames (ORFs), since ultimately it seems likely that the analysis of the function of a gene will depend on the characterization of its protein product. In our design, putative ORFs are initially cloned into a vector from which they can be easily transferred to multiple other plasmids for more specialized purposes, such as protein expression or mutant generation. Amplified ORFs flanked by bacteriophage lambda attB (bacterial attachment) sites are generated by a nested PCR protocol and cloned by lambda integrase-mediated recombination into an entry plasmid that contains the corresponding attP (phage attachment) sequences (attB-attP [BP] reaction). The result is a plasmid with two attL sites flanking the ORF of interest. Our entry plasmid is a modification of one commercialized in Gateway Technology by Invitrogen (http://www.invitrogen.com:80/content/sfs/manuals/gatewayman.pdf). We introduced an oriT sequence that allows the plasmid to be mobilized between Escherichia coli strains.

oriT is the basis of a procedure for recombining genes from these entry vectors into specialized destination plasmids by the lambda excisionase-mediated recombination reaction. As embodied in Gateway Technology, this second step is expensive and time-consuming, since it requires purified DNA substrates, recombination by enzymes in an in vitro recombination reaction, introduction of the recombined DNA into cells by transformation or electroporation, and screening for the correct recombinants. Using the oriT sequences on destination plasmids, we have developed a pentaparental mating procedure for shuttling genes from entry vectors to destination vectors in vivo. This mating strategy, which is adaptable to a 96-well format and is highly scalable, should allow the transfer of cloned ORFs en masse into various destination plasmids.

The destination plasmids used in Gateway Technology can be of several types. Currently available vectors can be used to express proteins in E. coli or tag the proteins in various ways to assist in purification or in monitoring gene expression. We have modified some of these procedures for use with S. meliloti and will report on these elsewhere.

One genetic manipulation that we considered very important, generating mutants that have defects in the identified ORFs, was not accessible with currently used techniques. A relatively simple and commonly used way for disrupting an ORF is to clone an internal fragment of the ORF onto a suicide plasmid and then select for integration of that plasmid into the bacterial chromosome by homologous recombination. This process leads to two partial copies of the ORF, one truncated at the N terminus and the other truncated at the C terminus. The resulting strain usually lacks a functional version of the ORF, although homologous recombination that reverses the plasmid integration can restore the original configuration. This internal-fragment strategy is not compatible with our desire to work with entire ORFs, since the integration of a complete ORF by homologous recombination does not lead to a truncated version of the ORF.

As an alternative, we constructed two distinct suicide vectors, each containing two copies of the yeast Flp recombinase target (FRT) sequences, which we integrated into the genome by gene-specific homologous recombination. We then generated deletion mutants through site-specific recombination reactions between direct repeats of the integrated FRT sequences. To demonstrate the usefulness of our overall scheme, we generated four deletion mutants of S. meliloti and assayed the phenotype of each mutant in relation to its symbiosis with alfalfa. The deletion mutations were each constructed within the denitrification cluster, a 60-gene cluster that is found on pSymA and that appears to be responsible for the conversion of nitrate to dinitrogen (denitrification). We showed that each of the four deletion mutants was able to form effective nodules on alfalfa roots and fix nitrogen at rates comparable to those of the wild type. We further studied for two of these mutants the process of nitrate-induced nodule senescence in alfalfa; the results of those studies will be reported elsewhere.

MATERIALS AND METHODS

Growth media, solutions, and buffers.

Luria-Bertani (LB) (23), minimal mannitol ammonium (MMNH4) (24), SOC (23), and yeast extract mannitol broth (YMB) (24) broth and agar were prepared as previously described. A 10% glycerol solution was prepared by bringing 100 ml of glycerol to 1 liter with deionized distilled water (ddH2O) and autoclaving. An 0.85% NaCl solution was prepared by adding 8.5 g of NaCl to 1 liter of ddH2O and autoclaving. Agarose gel running buffer, Tris-borate-EDTA buffer, and Tris-EDTA (TE) buffer were prepared as previously described (23). Stock solutions of ampicillin, gentamicin, isopropyl-β-d-thiogalactopyranoside (IPTG), kanamycin, penicillin, and spectinomycin were dissolved in ddH2O and filter sterilized. Chloramphenicol was dissolved directly in 100% ethanol, rifampin was dissolved in 100% methanol, and tetracycline was dissolved in 50% ethanol. Antibiotics and IPTG were added to the media as follows (a representative example for each antibiotic or IPTG is given): Cam25, chloramphenicol at 25 μg/ml; Gen20, gentamicin at 20 μg/ml; IPTG20, IPTG at 20 μg/ml; Kan75, kanamycin at 75 μg/ml; Rif100, rifampin at 100 μg/ml; Spc100, spectinomycin at 100 μg/ml; and Tet10, tetracycline at 10 μg/ml.

Bacterial strains and plasmids.

The E. coli and S. meliloti strains used in this study are summarized in Table Table1.1. The DNA vectors used or created in this work are summarized in Table Table2.2. Details of some of the procedures used to obtain these genetic materials follow.

TABLE 1.
E. coli and S. meliloti strains
TABLE 2.
Plasmid vectors and constructions

(i) pMK2010.

To insert the plasmid RP4 origin of transfer (oriTRP4) into the Gateway Technology entry plasmid, pDONR201, pBSL237 (2) was cut with restriction enzyme SphI to excise a fragment that contains oriTRP4 from the vector backbone. The restriction digest was fractionated by electrophoresis, and the oriTRP4 fragment was cut from the gel and purified. pDONR201 was cut with NspI (New England BioLabs, Beverly, Mass.), linearizing the plasmid and generating compatible cohesive ends. Shrimp alkaline phosphatase (U.S. Biochemical Corp., Cleveland, Ohio) was added to remove the terminal phosphates from pDONR201. Linearized pDONR201 and the oriTRP4 fragment were mixed and ligated overnight at room temperature. The resulting ligation mixture was transformed into strain DB3.1, and the sample was grown overnight on LB-Cam25 agar plates. Colonies were restreaked and screened by restriction digestion for the presence of pMK2010. The DNA sequences of the oriTRP4 fragment and adjacent regions then were determined to establish the precise sequence and orientation of the oriTRP4 cassette (Fig. (Fig.11).

FIG. 1.
Entry plasmid pMK2010. pMK2010 was cloned by insertion of oriTRP4 from pBSL237 into pDONR201. The DNA sequence is available as GenBank accession number ...

(ii) pMK2012 and pMK2013.

pPS854, which contains two FRT sequences in direct orientations (16), was cut at the single EcoRV site between these sequences, linearizing the plasmid and leaving blunt ends. Reading frame cassette B (Invitrogen), which contains lambda attP sequences flanking ccdB and a chloramphenicol resistance gene, was inserted into pPS854. Both orientations of the cassette were found, and the two resulting constructs were named pMK2012 and pMK2013 (Fig. (Fig.22).

FIG. 2.
Construction of deletion destination plasmids. Plasmids pMK2012 and pMK2013 were constructed as described in the text, cut with BglII, and ligated together. After HindIII digestion of the ligase reaction mixture (pMK2015) or PCR followed by HindIII digestion ...

(iii) pMK2014 and pMK2015.

To construct cassettes that contained the FRT sequences in both inverted orientations, pMK2012 and pMK2013 were cut with the endonuclease BglII, mixed together, and ligated. Four possible FRT orientations (relative to the chloramphenicol resistance and ccdB genes) were possible in this ligation. To construct pMK2014, a single PCR primer sequence that was outside the outwardly facing FRT sequence in pPS854 was used. This primer amplified the ligation products that contained both outwardly facing FRT sequences. After amplification, this PCR fragment was cut with HindIII and inserted into the HindIII site in pMB419. Since E. coli does not propagate direct inversions (11), the only amplified and clonable product is the one shown inserted into pMK2014. In theory, a similar strategy with a primer specific for the inwardly facing FRT sequence should have produced pMK2015, but this type of clone was not recovered. As an alternative, the initial ligation mixture described above was cut with HindIII and ligated to pMB419. Chloramphenicol-resistant transformants were screened by PCR with Taq polymerase and a primer specific for the inwardly facing FRT sequence. The orientation was confirmed by DNA sequencing.

(iv) pMK2016.

The deletion strategy described below requires two mobilizable plasmids that have no extended regions of homology. The mobilizing plasmids chosen, pRK2013 and pRK2073, can mobilize plasmids containing oriT sequences derived from plasmid RP4, RK2, or ColEI. To construct a destination plasmid with the ColE1 replicon and ColE1 origin of transfer (oriTColE1), the replicon and oriT fragment of pML21, a mini-ColEI plasmid (15), were separated from the kanamycin resistance region by using restriction enzymes EcoRI and KpnI. An aadA gene, which confers resistance to streptomycin and spectinomycin, was excised from pVO122 (3) by using EcoRI and KpnI and was ligated to the pML21 replicon to create pBH21. pBH21 was converted to a destination vector by excising the attR1-Camr-ccdB-attR2 cassette from pMK2014 by using HindIII and inserting this cassette into the unique HindIII site in pBH21. Both orientations of the cassette in pBH21 were obtained, and the one designated pMK2016 is shown in Fig. Fig.22.

(v) pMK2017.

To create a second plasmid for our deletion strategy, the attR1-Camr-ccdB-attR2 cassette from pMK2015 was excised by using KpnI and was inserted into the KpnI site in pKNOCK-Tc (1) (Fig. (Fig.22).

(vi) pBH474.

In order to catalyze recombination between FRT sequences, the Flp recombinase enzyme must be present. pTH474, a pBBR1-derived Genr plasmid, was used by Chain (7) in order to express Flp recombinase in S. meliloti. As part of our strategy, we wanted to eliminate the plasmid that donated Flp after excision had taken place. We inserted a sacB-sacR gene cassette, which encodes levansucrase and makes cells sensitive to sucrose, into pTH474. The cassette was excised from pCPP53 by using XmaI and was ligated to the XmaI site in linearized pTH474. The resulting DNA was transformed into strain DB3.1. Gentamicin-resistant, sucrose-sensitive colonies were recovered, and a plasmid from one of these was named pBH474.

(vii) DB3.1 λpir.

The ccdB gene carried on the destination plasmids is not toxic to strains, such as DB3.1, that carry certain gyrA mutations. Therefore, a λpir lysogen of DB3.1 was needed to propagate destination plasmids that use the R6K γ origin of replication. A 5-ml LB medium culture of S17-1 λpir was shaken overnight at 30°C, and an aliquot was centrifuged to pellet the cells. Several drops of chloroform were added to the supernatant, which contained λpir phage, and this solution was mixed for 5 min to kill the remaining S17-1 λpir cells. The aqueous fraction was passed through a Nalgene 0.2-μm-pore-size syringe filter. A 100-μl quantity of the sterile supernatant was added to 300 μl of an overnight culture of DB3.1, LB top agar was added, and this mixture was poured onto an LB medium plate. The plate was incubated at 37°C after the top agar had solidified. Bacteria were picked from the center of two of the largest plaques and streaked for isolation on an LB medium plate. Individual colonies were tested for their ability to maintain plasmids possessing a ccdB gene and an R6K γ origin of replication.

(viii) Rifampin-resistant E. coli strains.

Rifampin was a suitable resistance marker for recipient cells because no rifampin resistance gene is commonly used in other DNA cloning strategies. To select for rifampin resistance, 109 cells from a sensitive E. coli strain were plated on LB-Rif100 agar and incubated overnight at the appropriate temperature. Single colonies were isolated and routinely designated with an “R” immediately following the strain name.

Cloning into entry vectors. (i) Preparation of electrocompetent E. coli cells.

E. coli strains were inoculated into 5 ml of LB broth containing the appropriate antibiotic(s) and shaken overnight at 30°C. The culture was diluted to between 1:25 and 1:100 in LB broth and shaken (200 to 240 rpm) at 30°C. For most E. coli strains, the culture was harvested by centrifugation when the absorbance at 600 nm was between 0.8 and 1.2. The cells were washed successively in 1/5, 1/10, 1/20, and 1/40 the original culture volume of chilled 10% glycerol. The final pellet was resuspended in 1/250 the original culture volume of 10% glycerol. Aliquots of 45 to 90 μl were distributed to microcentrifuge tubes, quickly frozen in liquid nitrogen, and stored at −80°C. If used the same day, aliquots were kept on ice.

(ii) Electrotransformation of E. coli cells.

Aliquots of 45 to 90 μl of electrocompetent E. coli cells were removed from −80°C storage and thawed on ice. DNA (≤1 μg and ≤10 μl) was added to a 90-μl sample. The cell suspension and DNA were mixed and placed in a chilled electroporation cuvette with a 0.1-cm gap (Bio-Rad, Richmond, Calif.). The cuvette was pulsed at 1.5 to 1.8 kV in an electroporator (BTX TransPorator Plus). The cell suspension was immediately mixed with 160 to 950 μl of SOC medium and transferred to a 1.5-ml microcentrifuge tube. The tube was incubated without agitation at 37°C for 1 to 3 h. After incubation, aliquots of the cell suspension were spread on selective agar media and incubated at 37°C for 12 to 24 h.

(iii) BP Clonase reaction.

The BP Clonase (Invitrogen) reaction mediates recombination between the attB1 and attB2 sites flanking the PCR product and the attP1 and attP2 sites in the entry plasmid, pMK2010. The result is insertion of the PCR-amplified gene sequence into the entry plasmid. The following ingredients were mixed thoroughly in a 1.5-ml microcentrifuge tube: 2 μl of BP reaction buffer (Invitrogen), 4 μl of PCR product, 1 μl of pMK2010 (150 ng/μl), 2 μl of TE buffer or ddH2O, and 1 μl of BP Clonase enzyme mix. The tube was incubated at room temperature for 1 to 4 h. To remove proteins bound to the DNA, 1 μl of proteinase K (20 mg/ml) was added, and the tube was incubated at 37°C for 15 min. A total of 1 to 10 μl of this mixture was transferred to transformation-competent or electrocompetent DH5α cells. Following heat shock treatment or electroporation, cells were plated on LB-Kan75 agar and incubated overnight at 37°C.

(iv) PEG-Mg2+ purification of PCR products.

Polyethylene glycol (PEG) precipitation was used to remove low-molecular-weight PCR products that contained attB sequences not associated with the desired insert. A total of 50 μl of PCR product was added to 150 μl of TE buffer and mixed. A total of 100 μl of a 30% PEG 8000-Mg2+ solution (Invitrogen) was added, and the solution was mixed by vortexing. This mixture was centrifuged at ≥12,000 × g in a microcentrifuge at room temperature for at least 15 min. The supernatant was carefully removed, the tube was dried at 37°C for 10 to 15 min, and the pellet was resuspended in 30 to 50 μl of TE buffer. The purified PCR product was ready for use in the BP reaction.

In vivo transfer of cloned DNA to destination vectors.

For pentaparental matings, five strains in 5-ml cultures were shaken separately overnight at 30°C. DH5α(pXINT129), DH5α(pE entry vector), and HB101(pRK2013) were grown in LB-Kan75 medium. DB3.1(pMK2016) or DB3.1 λpir(pMK2017) was grown in LB-Cam50 medium. For transferring DNA to pMK2016, HB101R was used as the recipient strain. For transferring DNA to pMK2017, DH5αR λpir was used as the recipient strain. Both recipients were grown in LB medium. After overnight incubation, 1.5 ml of each of the five cultures was harvested and pelleted. Each pellet was washed twice in 1 ml of 0.85% NaCl. Each final pellet was resuspended in 50 to 500 μl of 0.85% NaCl. All five strains then were mixed together in a 1:1:1:1:1 ratio. A total of 50 to 200 μl of this mixture was spotted on an LB-IPTG20 agar plate and incubated for 12 to 18 h at 37°C. After incubation, a loopful of the spot was streaked for isolation on an LB-Spc100Rif100 agar plate (for pMK2016) or an LB-Tet10Rif100 agar plate (for pMK2017) and incubated overnight at 37°C.

Creation of strain Sm1021 deletion mutants. (i) Integrating the pD1 destination plasmid into the strain Sm1021 genome.

Three cultures were shaken overnight at 30°C. HB101(pRK2013) was grown in LB-Kan75 medium. HB101R derivatives containing pD1 destination vectors derived from pMK2016 were grown in LB-Spc100 medium. Sm1021 was grown in YMB medium. Samples of 1.5 ml of the overnight cultures were pelleted, and each pellet was washed twice in 1 ml of 0.85% NaCl and resuspended in 50 to 200 μl of 0.85% NaCl. The three strains then were mixed together in a 1:1:1 ratio. A total of 50 to 200 μl of this mixture was spotted on a YMB medium plate and incubated for 12 to 24 h at 30°C. A loopful was streaked on an MMNH4-Spc200 agar plate and incubated for 2 to 3 days at 30°C to isolate single colonies. Counterselection against the parental strains was done with the auxotrophic markers in HB101 and the spectinomycin sensitivity of Sm1021. The resulting colonies, referred to as Sm1021::pD1 strains, had a destination plasmid of the first type (pD1) inserted by homologous recombination at the site of the gene carried by this plasmid and were used in subsequent manipulations.

(ii) Integrating the pD2 destination plasmid into Sm1021::pD1 single integrants.

The manipulations described above were repeated, with the exception that DH5αR λpir strains carrying ORF-containing destination plasmids derived from pMK2017(pD2) were grown on LB-Tet10 agar and each of the Sm1021::pD1 single integrants was grown on YMB-Spc200 agar. The cultures were grown, washed, mixed, and incubated as described above, after which colonies were isolated by streaking on MMNH4-Tet1Spc200 agar plates and incubation for 2 to 3 days at 30°C. The resulting colonies, referred to as Sm1021::(pD1+pD2) strains, had a destination plasmid of the second type (pD2) inserted by homologous recombination at the site of the gene carried by this plasmid.

(iii) Creating the Sm1021 deletion mutants.

To create a deletion strain from an Sm1021::(pD1+pD2) strain, the Flp recombinase encoded on pBH474 must be introduced and the integrated structure must be resolved. A triparental mating with HB101(pRK2073), DB3.1(pBH474), and each Sm1021::(pD1+pD2) double integrant was established by growing each parent in selective media, concentrating the cells, and mating on solid YMB media as described above. After incubation for 12 to 24 h at 30°C, a loopful was streaked on a YMB-Gen20 agar plate, and single colonies were isolated after 2 to 3 days of incubation at 30°C. These colonies were restreaked on YMB-Gen20 agar plates and incubated for 3 days at 30°C. DB3.1(pBH474) was not able to grow on YMB-Gen20 agar plates, as the gentamicin resistance gene was effective in E. coli only on rich media at gentamicin concentrations of less than 10 μg/ml. These single colonies were replica plated on MMNH4-Tet1 agar and MMNH4-Spc200 agar plates to test for the excision of the pMK2016- and pMK2017-derived plasmids from the Sm1021 genome. Tets Spcs clones were streaked on YMB-5% sucrose agar plates and incubated for 3 days at 30°C to select for the loss of pBH474. Single colonies from these plates were screened for deletions. Primers flanking the putative deletions were used to amplify the targeted segment of DNA. The DNA sequence of this PCR product was determined in order to verify the deletion of the targeted DNA. We were also successful in carrying out this deletion procedure as a biparental mating with S17-1(pBH474) instead of HB101(pRK2073) and DB3.1(pBH474).

Alfalfa seed sterilization, planting, and growth.

Alfalfa (Medicago sativa) seeds (Ladak variety; lot A6-008; Bruce Seed Farm, Inc.) were washed and planted as described previously (24). Magenta plant boxes were each filled with 250 g of sand (Lane Mountain Company) and autoclaved to sterilize. Plant nutrient solution lacking any source of nitrogen was added (75 ml/box). A total of five to eight seedlings were planted per box, depending on the experiment. Each box was placed in a growth chamber kept at 23 to 25°C with continuous light. Inverted boxes were used to cover the plants, and the boxes were sealed until harvest.

S. meliloti inoculation of alfalfa plants.

An S. meliloti culture, shaken for 24 to 48 h at 30°C, was harvested, pelleted (4,000 rpm for 10 min), and washed twice in 0.85% NaCl (10 ml and then 5 ml). The final pellet was resuspended in 0.85% NaCl, and 100 μl of the cell suspension was used to inoculate each seedling on the same day as planting.

Acetylene reduction assay.

Alfalfa nodules were picked from the roots and weighed. A sample of 10 to 50 mg was placed in a 0-ml glass serum bottle. The bottle was sealed with a rubber septum, and 1 ml of air was replaced with acetylene. Ethylene production was measured by using a Shimadzu GC-8A gas chromatograph.

Nucleotide sequence accession numbers.

The sequence of pMK2010 has been assigned GenBank accession number AY423863. The sequence of pMK2016 has been assigned GenBank accession number AY423864. The sequence of pMK2017 has been assigned GenBank accession number AY423865.

RESULTS AND DISCUSSION

In order to exploit the wealth of bacterial DNA sequence information that is becoming available, improved methods for functional analysis are needed. In a consideration of the techniques that benefit most directly from the sequence, there are a number of distinct objectives that should be part of a complete characterization of the genes and their products. Drawbacks to working at the genomic level are the effort and cost involved in working on thousands of genes simultaneously. For example, at $0.30 per base, a set of 20-base primer sequences for PCR for the 6,200 ORFs in S. meliloti would cost about $75,000, and working with these sequences entails significant labor and associated expenses. Thus, a strategy that allows a single primer set to be used for multiple and various manipulations would have some significant advantages, particularly if it could be integrated with information from other techniques.

This single primer set could be used to generate PCR products that could be cloned into initial entry vectors by using lambda integrase-mediated recombination. Then, from these base entry vectors, gene sequences could be transferred to any number of destination plasmids, again by lambda integrase-mediated technology. This technology (Gateway Technology) allows orientation-specific cloning amenable to high-throughput procedures.

Nested PCR protocol generating BP reaction substrates. (i) Primer design.

The first step in cloning genes by using the Gateway Technology was to optimize a PCR protocol to generate PCR products that could be cloned into an entry vector by the BP reaction. To decrease the cost of DNA primers, we adapted a nested PCR protocol with gene-specific primers to amplify the gene target and secondary primers containing overlapping sequences to add the attB sites.

We focused on a subset of genes found within a 60-kb, ~60-gene cluster on pSymA that contains many genes predicted to be involved in denitrification, the conversion of nitrate to dinitrogen. We designed primer pairs for the fixS1, fixN1, nnrR, nirK, norD, and norE genes. Each gene-specific forward primer (primary forward primer) consisted of the 12 bases at the 3′ end of the attB1 sequence followed by 20 bases specific to the 20 bases immediately upstream of the putative start codon of the gene. Each gene-specific reverse primer (primary reverse primer) consisted of the 12 bases at the 3′ end of the attB2 sequence followed by 20 bases specific to the 19 terminal bases of the gene (including the stop codon) plus 1 base following the stop codon. The primary forward primers were designed such that the start codon would be in frame once the gene had been transferred to a destination vector created to generate N-terminal fusion proteins. In the primary reverse primers, the stop codon of each gene was changed to the amber-suppressible codon, TAG, to potentially allow an amber suppressor mutation to read through the stop codon and create C-terminal protein fusions (Fig. (Fig.3a).3a). The forward and reverse secondary primers contained the entire attB1 or attB2 sequences preceded by four guanidine nucleotides (Fig. (Fig.3b).3b). The primary forward and reverse primers were added to the initial reaction mixtures, and the secondary forward and reverse primers were added just before the second step of the nested protocol.

FIG. 3.
Cloning of genes by integrase recombination. (a) To generate a PCR product containing attB sites flanking gene X, a nested PCR protocol was used. The primary forward and reverse primers hybridize to gene X at their 3′ ends and contain 12-bp extensions ...

To optimize the nested PCR protocol, we limited the quantity of the primary primers, so that the generation of a substantial product depended on the addition of the secondary primers. We found the optimal final concentration of the primary primers to be between 31 and 63 fmol/μl, as determined by comparisons of the agarose gel band intensities between samples with and without secondary primers added. Using 50 fmol of primary primer/μl, we optimized the number of cycles to be run in each step of the nested protocol. Our final optimization used two steps of 20 cycles each, for a total of 40 cycles.

(ii) PEG-Mg2+ purification of PCR products.

When PCR products were used directly in the cloning reaction, up to 50% of the clones did not contain the desired gene but instead contained a DNA fragment ≤200 bp in length. Several attempts to sequence these insertions were unsuccessful, suggesting an aberrant DNA structure. Product literature from Invitrogen suggests that attB primers and/or primer dimers can efficiently clone into entry plasmids and recommends a PEG-Mg2+ purification procedure to eliminate PCR DNA shorter than 300 bp from recombination reactions. PEG precipitation improved the frequency of cloning of the desired DNA fragment into the entry plasmid, raising the success rate from 50% to over 83%. Unfortunately, a significant number of putative S. meliloti ORFs are between 200 and 400 bp in length.

Cloning of S. meliloti genes using the BP reaction.

The BP reaction was designed by Hartley et al. (13) to allow orientation-specific DNA cloning. To accomplish this, they developed cloning vectors with two distinct attB and attP sites, such that the attB1 site would recombine only with attP1 and the attB2 site would recombine only with attP2. The general strategy for creating entry vectors is diagrammed in Fig. Fig.3c.3c. Each PCR product, consisting of an ORF flanked by attB1 and attB2 sites, was mixed with our entry plasmid, pMK2010, and BP Clonase enzyme mix, which contains enzymes needed for site-specific recombination between attB and attP sites. After 1 h of incubation at room temperature, the product was electroporated into DH5α cells and plated on kanamycin agar. Kanamycin selects for the presence of the plasmid pMK2010 backbone, while the sensitivity of DH5α to the ccdB toxin gene located between the plasmid attP sites selects for replacement of the ccdB-Camr cassette with the PCR product. Resulting clones were screened for chloramphenicol sensitivity, and the recombinant plasmids were analyzed by restriction enzyme digestion.

The fixS1, fixN1, nnrR, nirK, norD, and norE genes were cloned into pMK2010 (Fig. (Fig.1)1) by using the BP reaction. The resulting entry vectors were named with the designation “pE” (for “entry plasmid”), followed by the location of the ORF on the S. meliloti genome (e.g., SMa1241 represents S. meliloti pSymA ORF 1241). Accordingly, these entry vectors were named pESMa1208 (fixS1), pESMa1220 (fixN1), pESMa1245 (nnrR), pESMa1250 (nirK), pESMa1269 (norD), and pESMa1279 (norE). Each clone was archived and stored in glycerol or dimethyl sulfoxide at −80°C.

We added 4 μl of PCR product to a 10-μl BP reaction mixture and then transformed 5 μl of this BP reaction mixture into DH5α cells by electroporation, which we found to be the most reliable method of producing transformants. When electroporation efficiencies for pUC19 were between 109 and 1011 transformants/μg of DNA, electroporation of the recombinase reaction products generally yielded between 1,000 and 5,000 transformants per electroporation. However, when electroporation efficiencies were below 109 transformants/μg of DNA, the number of BP reaction transformants per electroporation decreased, sometimes resulting in very few colonies. This number was obtained by simply counting kanamycin-resistant CFU per plate, so this measure does not reflect the number of clones per BP reaction carrying correct entry vectors.

In vivo transfer to destination plasmids.

For construction of deletion mutants of Sm1021, two deletion destination plasmids, pMK2016 and pMK2017, were created; these plasmids possess an attR1-ccdB-Camr-attR2 cassette, oriT, and yeast FRT sequences oriented as shown in Fig. Fig.2.2. A method of transferring cloned entry vector DNA into destination plasmids in vivo by using conjugation to bring the reaction components together was developed in order to circumvent several steps needed to recombine the ORFs into destination plasmids in vitro. We first created an entry plasmid, pMK2010, that was mobilizable so that it could be shuttled between E. coli strains with the aid of helper plasmids that provide needed transfer functions. This was done by inserting oriTRP4 obtained from plasmid pBSL237 into pDONR201. pMK2010 was able to move between E. coli strains in the presence of helper plasmids pRK2013 and pRK2073.

To facilitate integrase- and excisionase-mediated recombination in vivo, we used plasmid pXINT129 (21), which contains the lambda int and xis genes driven by the lac promoter and which expresses both the lambda integrase and the lambda excisionase proteins. lacI is also carried on the plasmid, so that Int and Xis can be induced by adding IPTG to the medium. It was shown previously that pXINT129 effectively catalyzes an exicision reaction between attR and attL sites (21).

In the pentaparental mating carried out as diagrammed in Fig. Fig.4,4, the five strains involved included strains carrying a helper plasmid (pRK2013 or pRK2073), pXINT129, an entry vector (pEGeneX), and a destination plasmid (pMK2016 or pMK2017) and a recipient rifampin-resistant strain. We hypothesize the following roles for each of the five strains. The strain containing the helper plasmid can donate this plasmid to any of the other strains in the mating mixture and can mobilize resident mobilizable plasmids. In the first phase of a successful mating, we propose that the helper plasmid moves into the strains that contain the entry and destination plasmids and allows these to be transferred to the strain that carries pXINT129 (not mobilizable). In the presence of Int and Xis produced from pXINT129, recombination transfers the cloned ORF DNA from the entry vector to the destination plasmid, replacing the ccdB gene on the destination plasmid. The destination vector, now containing the cloned DNA, can be mobilized into the final rifampin-resistant recipient without leading to cell killing. Selection for the antibiotic resistance carried on the destination plasmid and for rifampin resistance isolates cells of the final recipient that contain the destination vector (Fig. (Fig.4F).4F). The plausibility of this hypothesis is supported by the observation that omission of any one of the five strains eliminates the production of recombinants. Successful transfer occurred at a very high frequency, typically in the range of hundreds to thousands of successful transconjugants per 5-mm spot of mating mixture. A protocol has been developed that first spots a mixture of all strains except the one containing the entry plasmid on solid agar and then spots the strain containing the entry plasmid on this (data not shown). Recovery of the appropriate recombinants is efficient, and the protocol is suitable for a 96-well microtiter plate format.

FIG. 4.
Scheme for pentaparental mating events during in vivo construction of destination plasmid derivatives. The helper plasmid (pRK2013) (A) is first transferred (B) and established within each of the three strains shown (C). pRK2013 then allows the entry ...

(i) Transfer to pMK2016.

The fixN1, nirK, and norE genes were each transferred to pMK2016 by the in vivo pentaparental mating method with pRK2013 as the helper plasmid, since pRK2073 is spectinomycin resistant. The spectinomycin-rifampin selection was not sufficient to eliminate pMK2016-containing cells without single-colony isolation. Therefore, after mating and selection on LB-Spc100Rif100 medium plates, two or three individual colonies were picked and streaked for further isolation on LB-Spc100Rif100 medium plates. These deletion destination vectors were named with the designation “pD1′” (for “deletion plasmid 1”). Accordingly, these destination vectors were named pD1SMa1220 (fixN1), pD1SMa1250 (nirK), and pD1SMa1279 (norE).

(ii) Transfer to pMK2017.

The fixS1, nnrR, and norD genes were each transferred to the pKNOCK-Tc derivative plasmid, pMK2017, by the in vivo pentaparental mating method with pRK2013 as the helper plasmid. A one-time selection for tetracycline and rifampin resistance was sufficient to successfully obtain the correct destination vectors. These deletion destination vectors were named with the designation “pD2′” (for “deletion plasmid 2”). Accordingly, these destination vectors were named pD2SMa1208 (fixS1), pD2SMa1245 (nnrR), and pD2SMa1269 (norD).

Creating deletion mutants of Sm1021.

To implement our deletion strategy, pMK2016 and pMK2017 were each engineered to possess an origin of replication unable to function in S. meliloti and to contain strategically placed FRT sequences that bracket the ORF (Fig. (Fig.5).5). Independent of the original orientations of the two ORFs, the result of the two integration events leaves parallel FRT sequences at the boundary of the chromosomal construct. Flp-mediated recombination can delete the region between the two original ORFs and leave only a single FRT sequence flanked by attB sites in the chromosome.

FIG. 5.
Generation of deletion mutants of S. meliloti. A pD1GeneA destination vector derived from pMK2016 and containing gene A (A) is mated with Sm1021. Strains containing plasmids integrated into the genome by homologous recombination are selected with spectinomycin ...

pMK2016 was derived from pML21 and thus carries a ColE1 replication origin (oriVColE1) and oriTColE1, which are not similar enough to the replicons or oriT sequences of other plasmids to recombine efficiently in S. meliloti. To construct a version of pML21 that could be used together with our kanamycin resistance entry vector, pMK2010, the kanamycin resistance determinant on pML21 was replaced with the aadA gene from pVO122, which confers spectinomycin resistance. The attR1-Camr-ccdB-attR2 cassette from pMK2014 then was inserted to yield pMK2016, which has FRT sequences that face away from the cassette.

pMK2017 was derived from pKNOCK-Tc, which carries a replication protein-dependent origin of replication (oriVR6K), oriTRP4, and a tetracycline resistance gene. This plasmid must be propagated in a strain that contains the Pi replication protein of R6K. The attR1-Camr-ccdB-attR2 cassette from pMK2015, flanked by FRT sequences facing toward the insert, was ligated to pKNOCK-Tc to yield pMK2017. DB3.1 λpir, which is a derivative of DB3.1 lysogenized with a lambda bacteriophage carrying the pir gene, was used to propagate pMK2017.

Neither oriVColE1 of pMK2016 nor oriVR6K of pMK2017 can replicate in Sm1021; therefore, resistance genes carried on these plasmids will not be maintained in Sm1021 unless they are integrated into the genome. In our design, we expected this integration to occur by homologous recombination between the cloned DNA insert and the resident sequence. In order to target the second destination plasmid to the correct site in the genome, we wanted to minimize the homology between pMK2016 and pMK2017 so that very little or no homologous recombination would occur between these two plasmids. With the exception of the short regions of the FRT and attB sites, these plasmids share very little homology. pMK2016 and pMK2017 and their derivatives can be mobilized by either pRK2013 or pRK2073.

Integrating pD1 and pD2 destination vectors into the Sm1021 genome.

The pD1SMa1220, pD1SMa1250, and pD1SMa1279 destination vectors were integrated into the Sm1021 genome by mating each pD1 destination vector with Sm1021 and using spectinomycin resistance to select for homologous recombination between the plasmid-borne and the genomic copies of each gene (Fig. 5A and C). We refer to these strains as Sm1021::pD1 single integrants. A very small number of spectinomycin-resistant mutants of HB101(pRK2013) formed colonies on MMNH4-Spc200 medium plates. These mutants were easily distinguished because they formed visible single colonies after 24 h, while Sm1021::pD1 single integrant colonies were not visible for 48 to 72 h.

The pD2SMa1208, pD2SMa1245, and pD2SMa1269 plasmids were integrated into the genome of Sm1021::pD1 single integrants by mating each pD2 destination vector with the appropriate Sm1021::pD1 single integrant and selecting for homologous recombination between the plasmid-borne and the genomic copies of each gene (Fig. 5B to D) by placing the mating mixture on a minimal medium containing both spectinomycin and tetracycline. We refer to these strains as Sm1021::(pD1+pD2) double integrants. Background colonies from this mating were very rare.

From the Sm1021::(pD1+pD2) double integrants, we created four deletion mutants by using the method diagrammed in Fig. Fig.5.5. Chain, working with Turlough Finan, has developed a method in which Flp recombinase is used to excise segments of DNA from one of the megaplasmids in S. meliloti (7). Several plasmids with the potential to express Flp were investigated, and pTH474 was found to be best for this purpose. In the course of those investigations, a tester strain, RmH940, was generated; this strain is Lac and contains a Tn5 element possessing a Neor gene and E. coli lacZ flanked by direct FRT repeats. When Flp recombinase is expressed, the sequence between the FRT repeats is excised, and the resulting clones are Neos and Lac and produce white colonies on media that contain 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) (7). We constructed pBH474 by inserting a sacB-sacR cassette into pTH474 in order to be able to select for the loss of pBH474 by selecting for sucrose resistance. pBH474 was able to excise the Neor Lac determinant in RmH940 very efficiently (data not shown).

pBH474 was mated with Sm1021::(pD1+pD2) double integrants, and the samples were streaked on YMB-Gen20 medium plates and incubated for 3 days at 30°C. Single colonies were picked, streaked on YMB-Gen20 medium plates, and incubated for 3 days at 30°C. We found that the additional purification and incubation time was necessary for the Flp recombinase to excise the two plasmid sequences and the genomic DNA (Fig. (Fig.5E)5E) and to segregate isolates with deletions. Following the second incubation period, single colonies were screened for sensitivity to tetracycline and spectinomycin. More than 75% of the resulting colonies were sensitive to both antibiotics.

Interestingly, selection for the presence of pBH474 in Sm1021::(pD1+pD2) double integrants was effective on YMB-Gen20 medium but not on MMNH4-Gen20 medium because, for unknown reasons, the strain was already gentamicin resistant on the minimal medium. Also, helper plasmid pRK2013 could not be used, as it could mate with DB3.1(pBH474), allowing resistance to gentamicin on YMB-Gen20 medium plates. Background colonies of DB3.1(pBH474) arose in the matings, but these could be differentiated easily from Sm1021 recipients. Mutant DB3.1(pBH474) single colonies were visible in 24 h, while Sm1021 recipients were not visible until 48 to 72 h.

Once deletions were made, we cured the mutants of plasmid pBH474 by growing them on YMB-5% sucrose medium plates. The resulting colonies were screened for gentamicin sensitivity. Four deletion mutants were generated and confirmed by PCR and DNA sequencing. The deletion mutants were named Sm1021::SMa1208ΔSMa1220, Sm1021::SMa1245ΔSMa1250, Sm1021::SMa1250ΔSMa1269, and Sm1021::SMa1269ΔSMa1279, indicating that, for each mutant, the entire DNA sequence between both named genes was deleted.

Effective nodulation of alfalfa by Sm1021 mutants.

In the manipulations described above, we generated four mutants with deletions in the denitrification cluster. Sm1021::SMa1208ΔSMa1220 lacks the fixI1, fixH, fixG, fixP1, fixQ1, and fixO1 genes; Sm1021::SMa1245ΔSMa1250 lacks the nirV gene; Sm1021::SMa1250Δ1269 lacks 10 putative ORFs between the nirK and norD genes; and Sm1021::SMa1269ΔSMa1279 lacks the norQ, norB, and norC genes. Sm1021::SMa1245ΔSMa1250 lacked nitrite reductase activity and accumulated nitrite in the presence of nitrate (data not shown). To demonstrate the utility of our deletion strategy in answering biological questions, we examined the ability of each mutant to form effective nodules on alfalfa (i.e., nodules actively fixing nitrogen).

Alfalfa seedlings were inoculated with each of the four mutant strains and wild-type Sm1021 in the absence of any nitrogen source. The alfalfa seedlings were grown for 35 days and harvested. The nodules from the mutants showed no significant differences (size, shape, number, and color) from the wild-type nodules. Acetylene reduction rates for the mutants also showed no significant differences from those for the wild type. The effective phenotypes of Sm1021::SMa1245ΔSMa1250, Sm1021::SMa1250Δ1269, and Sm1021::SMa1269ΔSMa1279 were not a surprise, since denitrification, although common in S. meliloti (20), is not a universal feature of this species. On the other hand, the Fix+ phenotype of Sm1021::SMa1208ΔSMa1220 was unexpected. There are additional genes in Sm1021 that are similar to fixI1, fixP1, fixQ1, and fixO1, and transposon insertions into the latter three of these do not alter effectiveness in a wild-type background because of functional complementation from other regions of pSymA (22). On the other hand, transposon insertions into fixG, fixH, and fixI1 lead to a Fix phenotype (17). FixG is predicted to be an iron sulfur membrane protein, and FixH is predicted to have a transmembrane region. The effectiveness of the Sm1021::SMa1208ΔSMa1220 mutant suggests that proteins expressed elsewhere from pSymA are functionally similar to FixG and FixH, but there are no other proteins in S. meliloti that have obvious sequence similarity. Reconciling our deletion results with the transposon results cited above may also require that, for fixGHI1 mutations to cripple effectiveness, some specific interactions with the other fix genes in this region may be needed.

Conclusions.

As the sequences of entire genomes have become available, the need for tools to investigate the functional aspects of individual genes within genomes has arisen. Eukaryotic functional genomic techniques have done well at meeting this need for selected organisms, although the efforts have been expensive. The techniques that we have outlined here should significantly increase the capabilities of genome-level experimentation with S. meliloti and possibly other prokaryotic organisms. We are currently cloning all of the identified ORFs of S. meliloti into pMK2010 in order to implement a strategy for genomic analysis based on the methods described here. Similar approaches with other bacteria may also be useful for designing a versatile platform for genetic studies at a reasonable cost.

One important use of genomic sequences is to help in creating mutants. We have demonstrated a method for mutagenesis that starts with cloned ORFs and provides an alternative to methods that clone internal fragments and generate mutants by homologous recombination. This method can be integrated into a more general cloning strategy and thus provide a way to integrate a deletion mutagenesis objective into a high-throughput genomic resource that starts with a single set of primers for each gene. The deletion mutations have some direct advantages in that they are not subject to reversion and the end points can be chosen as desired, making it possible to delete blocks of genes in addition to single genes. However, one difficulty with these deletions is that they extend to the ends of the ORFs adjacent to the ORF(s) to be deleted and thus have the potential to remove control sequences at the ends of these adjacent ORFs. This feature is a direct consequence of the strategy for choosing primers. The largest deletion that we made in this study was ~9 kb, but there is no reason to think that larger deletions cannot be constructed by Flp-mediated recombination.

The use of a variety of destination plasmids to accomplish various goals would seem to simply move the difficulty of large-scale manipulation of genomic sequences to a different set of in vitro reactions. However, we have also demonstrated here that by incorporating mobilization functions into the entry and destination plasmids, it is possible to create the required constructions in vivo at high efficiency. This method is rapid and inexpensive, since integrase reactions are very specific and since no DNA preparation or in vitro manipulation is required. The use of pMK2010 (or another plasmid that contains an oriT sequence) for the initial cloning is a minor procedural modification that enables subsequent in vivo recombinations to be carried out and appears to have no significant drawback. Incorporating FRT sequences into destination plasmids in order to construct deletions clearly has a more specialized purpose, but we have incorporated this strategy into plasmids to be used to construct reporter fusions and have found that it works very well.

While this report has demonstrated the utility of the Flp-FRT deletion strategy only for S. meliloti, Flp-FRT strategies have been used for other organisms (16). While it may take some effort to obtain Flp expression at appropriate levels, it seems likely that the deletion strategy presented here will be more generally useful. Our specific strategy should be useful for any organism that can conjugate with E. coli, where integration into the chromosome by homologous recombination can occur, and where Flp-FRT site-specific recombination will work. Such organisms would include the large number of bacteria that can mate with incP-1 plasmids (2, 3) and perhaps some lower eukaryotes, such as yeasts (14).

Acknowledgments

We thank Mikhail Alexeyev, Herbert Schweizer, Gregory Phillips, and, especially, Turlough Finan for genetic materials that served as starting points for the plasmid constructions and recombination procedures described in this work.

This research was supported by the Agricultural Research Center at Washington State University and by grants from the U.S. Department of Energy Energy Biosciences Program and the U.S. National Science Foundation Microbial Genetics Program.

REFERENCES

1. Alexeyev, M. F. 1999. The pKNOCK series of broad-host-range mobilizable suicide vectors for gene knockout and targeted DNA insertion into the chromosome of gram-negative bacteria. BioTechniques 26:824-826, 828. [PubMed]
2. Alexeyev, M. F., and I. N. Shokolenko. 1995. RP4 oriT and RP4 oriT-R6K oriV DNA cassettes for construction of specialized vectors. BioTechniques 19:22-24, 26. [PubMed]
3. Barnett, M. J., V. Oke, and S. R. Long. 2000. New genetic tools for use in the Rhizobiaceae and other bacteria. BioTechniques 29:240-242, 244-245. [PubMed]
4. Bernard, P., K. E. Kezdy, L. van Melderen, J. Steyaert, L. Wyns, M. L. Pato, P. N. Higgins, and M. Couturier. 1993. The F plasmid CcdB protein induces efficient ATP-dependent DNA cleavage by gyrase. J. Mol. Biol. 234:534-541. [PubMed]
5. Boyer, H. W., and D. Roulland-Dussoix. 1969. A complementation analysis of the restriction and modification of DNA in Escherichia coli. J. Mol. Biol. 41:459-472. [PubMed]
6. Cabanes, D., P. Boistard, and J. Batut. 2000. Identification of Sinorhizobium meliloti genes regulated during symbiosis. J. Bacteriol. 182:3632-3637. [PMC free article] [PubMed]
7. Chain, P. S. G. 1998. Development of an in vivo DNA cloning procedure for bacteria. M.S. thesis. McMaster University, Hamilton, Ontario, Canada.
8. 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. [PMC free article] [PubMed]
9. Galibert, F., T. M. Finan, S. R. Long, A. Puhler, P. Abola, F. Ampe, F. Barloy-Hubler, M. J. Barnett, A. Becker, P. Boistard, G. Bothe, M. Boutry, L. Bower, J. Buhrmester, E. Cadieu, D. Capela, P. Chain, A. Cowie, R. W. Davis, S. Dreano, N. A. Federspiel, R. F. Fisher, S. Gloux, T. Godrie, A. Goffeau, B. Golding, J. Gouzy, M. Gurjal, I. Hernandez-Lucas, A. Hong, L. Huizar, R. W. Hyman, T. Jones, D. Kahn, M. L. Kahn, S. Kalman, D. H. Keating, E. Kiss, C. Komp, V. Lelaure, D. Masuy, C. Palm, M. C. Peck, T. M. Pohl, D. Portetelle, B. Purnelle, U. Ramsperger, R. Surzycki, P. Thebault, M. Vandebol, F. J. Vorholter, S. Weidner, D. H. Wells, K. Wong, K. C. Yeh, and J. Batut. 2001. The composite genome of the legume symbiont Sinorhizobium meliloti. Science 293:668-672. [PubMed]
10. Geurts, R., and H. Franssen. 1996. Signal transduction in Rhizobium-induced nodule formation. Plant Physiol. 112:447-453. [PMC free article] [PubMed]
11. Hagan, C. E., and G. J. Warren. 1982. Lethality of palindromic DNA and its use in selection of recombinant plasmids. Gene 19:147-151. [PubMed]
12. Hanahan, D. 1983. Studies on transformation of Escherichia coli with plasmids. J. Mol. Biol. 166:557-580. [PubMed]
13. Hartley, J. L., G. F. Temple, and M. A. Brasch. 2000. DNA cloning using in vitro site-specific recombination. Genome Res. 10:1788-1795. [PMC free article] [PubMed]
14. Heinemann, J. A., and G. F. Sprague, Jr. 1989. Bacterial conjugative plasmids mobilize DNA transfer between bacteria and yeast. Nature 340:205-209. [PubMed]
15. Hershfield, V., H. W. Boyer, L. Chow, and D. R. Helinski. 1976. Characterization of a mini-ColE1 plasmid. J. Bacteriol. 126:447-453. [PMC free article] [PubMed]
16. Hoang, T. T., R. R. Karkoff-Schweizer, A. J. Kutchma, and H. P. Schweizer. 1998. A broad-host-range Flp-FRT recombination system for site-specific excision of chromosomally-located DNA sequences: application for isolation of unmarked Pseudomonas aeruginosa mutants. Gene 212:77-86. [PubMed]
17. Kahn, D., M. David, O. Domergue, M. L. Daveran, J. Ghai, P. R. Hirsch, and J. Batut. 1989. Rhizobium meliloti fixGHI sequence predicts involvement of a specific cation pump in symbiotic nitrogen fixation. J. Bacteriol. 171:929-939. [PMC free article] [PubMed]
18. Leigh, J. A., E. R. Signer, and G. C. Walker. 1985. Exopolysaccharide-deficient mutants of Rhizobium meliloti that form ineffective nodules. Proc. Natl. Acad. Sci. USA 82:6231-6235. [PMC free article] [PubMed]
19. Leong, S. A., G. S. Ditta, and D. R. Helinski. 1982. Heme biosynthesis in Rhizobium. Identification of a cloned gene coding for delta-aminolevulinic acid synthetase from Rhizobium meliloti. J. Biol. Chem. 257:8724-8730. [PubMed]
20. O'Hara, G. W., R. M. Daniel, and K. W. Steele. 1983. Effect of oxygen on the synthesis, activity and breakdown of the Rhizobium denitrification system. J. Gen. Microbiol. 129:2405-2412.
21. Platt, R., C. Drescher, S. K. Park, and G. J. Phillips. 2000. Genetic system for reversible integration of DNA constructs and lacZ gene fusions into the Escherichia coli chromosome. Plasmid 43:12-23. [PubMed]
22. Renalier M. H., J. Batut, J. Ghai, B. Terzaghi, M. Gherardi, M. David, A. M. Garnerone, J. Vasse, G. Truchet, T. Huguet, and P. Boistard. 1987. A new symbiotic cluster on the pSym megaplasmid of Rhizobium meliloti 2011 carries a functional fix gene repeat and a nod locus. J. Bacteriol. 169:2231-2238. [PMC free article] [PubMed]
23. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
24. Somerville, J. E., and M. L. Kahn. 1983. Cloning of the glutamine synthetase I gene from Rhizobium meliloti. J. Bacteriol. 156:168-176. [PMC free article] [PubMed]
25. van Rhijn, P., and J. Vanderleyden. 1995. The Rhizobium-plant symbiosis. Microbiol. Rev. 59:124-142. [PMC free article] [PubMed]
26. Yanisch-Perron, C., J. Vieira, and J. Messing. 1985. Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13mp18 and pUC19 vectors. Gene 33:103-119. [PubMed]

Articles from Applied and Environmental Microbiology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...