Rhizobium (Sinorhizobium) meliloti phn Genes: Characterization and Identification of Their Protein Products

Abstract

In Escherichia coli, the phn operon encodes proteins responsible for the uptake and breakdown of phosphonates. The C-P (carbon-phosphorus) lyase enzyme encoded by this operon which catalyzes the cleavage of C-P bonds in phosphonates has been recalcitrant to biochemical characterization. To advance the understanding of this enzyme, we have cloned DNA from Rhizobium (Sinorhizobium) meliloti that contains homologues of the E. coli phnG, -H, -I, -J, and -K genes. We demonstrated by insertional mutagenesis that the operon from which this DNA is derived encodes the R. meliloti C-P lyase. Furthermore, the phenotype of this phn mutant shows that the C-P lyase has a broad substrate specificity and that the organism has another enzyme that degrades aminoethylphosphonate. A comparison of the R. meliloti and E. coli phn genes and their predicted products gave new information about C-P lyase. The putative R. meliloti PhnG, PhnH, and PhnK proteins were overexpressed and used to make polyclonal antibodies. Proteins of the correct molecular weight that react with these antibodies are expressed by R. meliloti grown with phosphonates as sole phosphorus sources. This is the first in vivo demonstration of the existence of these hitherto hypothetical Phn proteins.

Phosphonates are organophosphorus compounds containing the chemically inert carbon-phosphorus (C-P) bond. Examples of naturally occurring phosphonates include phosphoenolpyruvate, 2-aminoethylphosphonate (2-AEP), and phosphonoacetate (PA) (16). In addition to these natural compounds, man-made phosphonates are now entering the environment in significant quantities (7). The ability to degrade phosphonates is relatively widespread, occurring in gram-positive (22, 39) and gram-negative bacteria (8, 39) as well as in fungi (20). Three classes of enzyme capable of breaking the C-P bond of phosphonates are known: PA hydrolase, an enzyme specific for PA breakdown (27, 30); phosphonatase, which specifically degrades 2-AEP (22, 24); and C-P lyase, which cleaves the C-P bond in a broad spectrum of phosphonates (10). C-P lyase activity can be detected in whole organisms; however, it has never been convincingly assayed in cell extracts (43), and this has limited attempts to understand the mechanism of the enzyme, which has been suggested to involve a redox-dependent free radical mechanism (10). The uptake and breakdown of phosphonates in Escherichia coli is, however, well characterized genetically (4). The phn gene cluster consists of 17 genes (phnA to -Q), of which phnC to -P appear to be required for phosphonate uptake and breakdown (33). Mutagenesis of the phn gene cluster revealed that phnCDE encode a phosphonate transporter, phnF and phnO may have regulatory functions, phnG to -M are likely to be components of the C-P lyase, and phnN and phnP are probably accessory proteins (34). To broaden knowledge of C-P lyase, we chose to work with Rhizobium (Sinorhizobium [6]) meliloti because (i) it contains a C-P lyase able to degrade the important herbicide N-phosphonomethyl-glycine (25), (ii) some phn genes in this organism have been sequenced (28), and (iii) a phosphate/phosphonate transporter, encoded by phoCDET, is needed for the successful formation of a symbiotic association between this bacterium and alfalfa (Medicago sativa) (1).

In this study, we cloned DNA encoding the putative phn genes from R. meliloti, constructed a mutation in the genome which establishes a role for these genes in phosphonate degradation, and demonstrated for the first time that proteins deduced to be encoded by phn genes are produced in vivo by organisms growing with phosphonates as the sole phosphorus sources.

MATERIALS AND METHODS

Sources of reagents.

Methylphosphonate (98%), ethylphosphonate (97%), propylphosphonate (95%), tert-butylphosphonate (98%), aminomethylphosphonate (99%), 3-aminopropylphosphonate, glyphosate (phosphonomethylglycine) (96%), phenylphosphonate (98%), and 4-aminobenzylphosphonate (97%) were obtained from Aldrich Chemical Company (Dorset, England). 2-AEP (99%), glycerol-3-phosphate (95%), and PA (98%) came from Sigma-Aldrich Ltd. (Dorset, England). Benzylphosphonate (97%) was from Lancaster Synthesis Ltd. (Lancashire, England).

Bacterial strains and growth conditions.

Strains used in this work are described in Table 1. R. meliloti was grown at 30°C either on TY (3) with 6 mM CaCl₂ or on acid minimal salts (36) modified by increasing the CaCl₂ concentration to 1.2 mM and adding nicotinic acid (1 mg/liter); the carbon source was 50 mM succinate, and phosphorus sources were provided at 0.5 mM unless otherwise stated. To make solid acid minimal salts medium containing a low level of inorganic phosphate, the medium was made double strength and added to a molten solution of 1.8% (wt/vol) agarose. E. coli was grown on LB (35). Antibiotics for R. meliloti were added at 500 (streptomycin), 20 (spectinomycin), and 10 (gentamicin) μg/ml; those for E. coli: were added at 100 (ampicillin), 20 (kanamycin), and 5 (gentamicin) μg/ml.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Relevant characteristics	Reference and/ or source
Strains
R. meliloti
Rm1021	SU47 str-21	32
Pn1	Rm1021 Δ(phnH–J):Ω Spr	This study
E. coli
BL21(DE3)	E. coli B F−dcm ompThsdS (rB−mB−) gal λ(DE3) [Cam]	41, Stratagene
DH5α	supE44 ΔlacU169 (φ80 lacZΔM15) hsdR1 recA1 endA1 gyrA96 thi-1 relA1	13, GIBCO BRL
GM1674	F′ proAB+ lacIqZΔM15 Δ(pro-lac) XIII dam-J dcm-6 thi-1 tsx63 λ−	17
Plasmids
pHP45Ω	pBR322 derivative containing Smr-Spr region flanked by Ω repeats	37
pJQ200SK	Mobilizable suicide vector with p15A origin of replication; contains RP4 origin of transfer and sacB from B. subtilis for counterselection; Gmr	38
pRK2013	ColE1 replicon with RK2 tra genes; helper plasmid used for mobilizing P- and Q-group plasmids; Nmr Kmr	9
pRSETB	ColE1 expression vector T7 promoter; Ampr	19, Invitrogen
pRMP1	pTZ19R carrying 3.9-kb XbaI/BamHI fragment containing PCR-amplified R. meliloti phnG–K	This study
pRMP2	Same as pRMP1 but with the 3.9-kb fragment cloned in pTZ18R	This study
pRMP3	pTZ19R carrying the same fragment as pRMP1+2 except with a 1.8-kb deletion from ClaI to NdeI and the deleted sequence replaced by the Smr-Spr fragment from pHP45Ω	This study
pRMP4	pJQ200SK carrying the 4.1-kb insert from pRUPn2 cloned into the XbaI site of the MCSa	This study
pTHGB	pRSETB carrying R. meliloti PCR-amplified phnG cloned into the BglII/HindIII site of the MCS; this fuses PhnG in frame to a His tag	This study
pTHHB	pRSETB carrying R. meliloti PCR-amplified phnH cloned into the BglII/HindIII site of the MCS; this fuses PhnH in frame to a His-tag	This study
pTHIB	pRSETB carrying R. meliloti PCR-amplified phnI cloned into the BglII/HindIII site of the MCS; this fuses PhnI in frame to a His tag	This study
pTHJB	pRSETB carrying R. meliloti PCR-amplified phnJ cloned into the BglII/HindIII site of the MCS; this fuses PhnJ in frame to a His tag	This study
pTHKB	pRSETB carrying R. meliloti PCR-amplified phnK cloned into the BglII/HindIII site of the MCS; this fuses PhnK in frame to a His tag	This study
pTHGC	pRSETB carrying R. meliloti PCR-amplified phnG cloned into the NdeI/HindIII site of the MCS; this does not fuse PhnG to the His tag	This study
pTHHC	pRSETB carrying R. meliloti PCR-amplified phnH cloned into the NdeI/HindIII site of the MCS; this does not fuse PhnH to the His tag	This study
pTHIC	pRSETB carrying R. meliloti PCR-amplified phnI cloned into the NdeI/HindIII site of the MCS; this does not fuse PhnI to the His tag	This study
pTHJC	pRSETB carrying R. meliloti PCR-amplified phnJ cloned into the NdeI/HindIII site of the MCS; this does not fuse PhnJ to the His tag	This study
pTHKC	pRSETB carrying R. meliloti PCR-amplified phnK cloned into the NdeI/HindIII site of the MCS; this does not fuse PhnK to the His tag	This study
pTZ19R/18R	Phagemid, pUC replicon carrying f1(−) origin and T7 promoter. pTZ19R and -18R have pUC19 and -18 polylinkers, respectively	31, Pharmacia

^aMCS, multiple cloning site. 

Genetic and DNA manipulations and sequencing.

All routine DNA manipulation and analysis were done as described in reference 40. Southern transfer to Amersham Hybond N nylon membrane and detection of hybrids by using the Boehringer Mannheim digoxigenin chemiluminescence kit were performed according to the manufacturer’s instructions. Probes were made by PCR amplification of plasmid inserts with digoxigenin-labeled dUTP in the mixture. Genetic conjugations were done by triparental matings according to the method of Figurski and Helinski (9), with E. coli DH5α as the donor strain and DH5α(pRK2013) used for the transfer functions. Transconjugants were selected on TY agar containing streptomycin and spectinomycin.

Nucleotide sequences were obtained by automated sequencing using a Pharmacia ALF express DNA sequencer. The sequencing reactions were done with an Amersham Thermosequenase kit according to the manufacturer’s instructions with Cy5-labeled primers.

PCR.

Oligonucleotide primers are described in Table 2. Reaction mixtures contained primers (0.2 μM), deoxynucleoside triphosphates (0.25 μM), MgCl₂ (2 mM), target DNA (0.5 μg), dimethyl sulfoxide (10%, vol/vol) 1× OptiPerform buffer (Bioline UK Ltd.), and 2 U of Bio-X-Act DNA polymerase (Bioline UK Ltd.) in a final volume of 50 μl. The mixture was overlaid with mineral oil and subjected to 30 cycles of 95°C for 30 s, 60°C for 30 s, and 68°C for 1 min/expected kb of product.

TABLE 2.

Oligonucleotide primers used in this work

Primer	DNA sequence (5′ to 3′)a	Location (nt)b
RmPhn1	CCTGTCATTTGCGAAGGTTATGG	221–243
RmPhn2	GGGGAGTTGGCATGGTCTTTAC	4186–4166
PhnGF	ATGATGGACGCAGCGAAGACAAGTGACGATG	294–324
PhnGR	ATCCTCTCCGCGGACCATGGTG	762–741
PhnHF	ATGACCGCCCAATCGCAAATCTATAGCGGC	764–793
PhnHR	CTCCCGCTCGAGTTTGGTGGTGC	1363–1341
PhnIF	ATGTATGTAGCCGTCAAGGGCGGAGAAGCC	1371–1400
PhnIR	TTCGGCGGCCTCCTTCATGCC	2474–2455
PhnJF	ATGAACGACCTCGCAACTTACAACTTCGC	2474–2502
PhnJiF	TCCAGCTCTTCGGCGCCGGCCGGGAGAAGC	3058–3078
PhnJiR	CCGGCGCCGAAGAGCTGGAGGGCCTCCGACATGTGCAT	3097–3059
PhnJR	TTCTGCAGCCTCCTGGCTTTCTTTCGC	3364–3339
PhnKF	ATGAGCGCCGTGCCGCTTCTG	3364–3384
PhnKR	CACCTGAAGAATCGAGGAAACGAGGAGC	4137–4111

^aThe EarI restriction site in phnJ is underlined; a silent mutation (to remove the NdeI site from phnJ) is in boldface. 

^bOn the existing R. meliloti phn sequence (28). 

Sequence analysis.

Use was made of the Genetics Computer Group (GCG) Wisconsin Package, release 8.1 (11). DNA sequences were assembled by using the GelAssemble package. Scans for expressed genes were done with TestCode and CodonFrequency, using an R. meliloti codon preference table based on GenBank release 104 from the codon usage database at the National Institute of Agrobiological Resources, Tsukuba, Japan (http://www.dna.affrc.go.jp/). Homology searches were done with the GCG implementation of FASTA and with BLAST (12) implemented at the National Biotechnology Information Center, using the default settings of both algorithms. Scans for sequence motifs were done with the scan ProSite tool at the Swiss Institute of Bioinformatics (http: //www.expasy.ch/) and GeneFind (44) (http://athena.uthct.edu/cgi-bin/genefind.pl). Multiple sequence alignments were done with Pileup (GCG), Multalin (5), and ClustalW (14).

Cloning and overexpression of putative phn genes of R. meliloti.

PCR primers (Table 2) were designed to amplify the nucleotide sequences of the genes encoding the putative Phn proteins. Restriction sites were added at the 5′ end of each primer for cloning purposes. Individual genes were amplified by PCR, and the resultant products were cloned into the pRSETB expression vector (Table 1). Two expression constructs were made for each putative gene. First, translational fusions between the N-terminal coding sequence of each gene product and the polyhistidine tag coding sequence of pRSETB were constructed. The second set of constructs fused the ATG start codon of the putative phn genes and the NdeI site of the pRSETB vector, giving plasmids potentially capable of expressing native gene products. However, the putative phnJ gene contained an internal NdeI site, which was removed by PCR as follows. A 5′ fragment of the gene was amplified by using primers PhnJF and PhnJiR, which contains an A-to-G substitution that removes the NdeI site without altering the predicted protein product. An overlapping 3′ fragment was amplified by using the primers PhnJiF and PhnJR (Table 2; Fig. 1). The two PCR products were ligated after digestion with EarI, which cuts at a unique restriction site in the gene. The ligated product was then cloned into pRSETB as described above. Both sets of constructs were electroporated into E. coli DH5α and E. coli BL21(DE3). The E. coli BL21(DE3) clones were subsequently used in overexpression experiments. Cells were grown overnight at 37°C with shaking in LB medium (10 ml) supplemented with ampicillin (100 μg/ml). Overnight cultures were used to inoculate fresh media, which were grown at 37°C to an optical density at 600 nm of 0.4 to 0.7. Isopropyl-β-d-thiogalactopyranoside (IPTG; 1 mM, final concentration) was added to the cultures, which were grown for a further 2 h prior to harvesting.

FIG. 1.

Inferred physical map of the DNA sequence of the R. meliloti phn gene cluster (28) showing locations of the PCR primers used in this study (small arrows above the sequence) and the predicted protein products. The 2-kb Sp^r-Sm^r fragment from plasmid pHP45Ω (37) was inserted into the DNA cloned in pRMP1 as a SmaI fragment into the ClaI and NdeI sites blunted with the Klenow fragment of DNA polymerase I (40). This construct was inserted into the XbaI site of pJQ200SK to give pRMP4, which was conjugated into R. meliloti to gave the mutant strain Pn1.

Analytical methods.

Cell extracts were prepared by harvesting early- to mid-log-phase cells of R. meliloti grown on various phosphorus sources. Cells were washed twice in 50 mM Tris-HCl buffer (pH 7.5), resuspended in the same buffer to 1/10 the original culture volume, and sonicated on ice. Cell debris was removed by centrifugation at 3,000 × g for 15 min at 4°C, and the extracts were stored at −20°C until required. The protein content of cell extracts was estimated by the method of Lowry et al. (26), using bovine serum albumin as a standard. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis (21) was carried out on 12% (wt/vol) polyacrylamide gels. Purification of His-tagged phn gene products was carried out by affinity purification using a 1-ml HiTrap chelating column (Pharmacia Biotech). Extracts from postinduced E. coli BL21(DE3) cells were prepared by harvesting and resuspension to one-fifth the original culture volume in guanidinium lysis buffer (20 mM sodium phosphate, 0.5 M sodium chloride, 6 M guanidine hydrochloride [pH 7.8]), rocking for 5 to 10 min at room temperature, and sonication on ice. Cell debris was removed by centrifugation at 3,000 × g for 15 min at 4°C, and extracts were stored at −20°C until required. HiTrap columns were prepared by washing with 5 column volumes (CV) of distilled water to remove the 20% ethanol storage buffer. The columns were primed by addition of 1 ml of 0.1M NiCl₂ · 6H₂O and then equilibrated with 5 CV of denaturing binding buffer (20 mM sodium phosphate, 0.5 M sodium chloride, 8 M urea [pH 7.8]); after loading of cell extracts, the columns were washed with a further 5 CV of denaturing binding buffer. The columns were then washed with 5 CV of denaturing wash buffer (20 mM sodium phosphate, 0.5 M sodium chloride, 8 M urea [pH 5.3]) to remove contaminating nonrecombinant proteins. Recombinant proteins were eluted from the columns with 5 CV of denaturing elution buffer (20 mM sodium phosphate, 0.5 M sodium chloride, 8 M urea [pH 4.0]). The purified proteins were then used to raise polyclonal antibodies in rabbits.

Western immunoblotting was carried out essentially as described in reference 40 except that blotted nitrocellulose filters were blocked overnight with blocking solution (50 mM Tris-HCl, 150 mM sodium chloride, 0.2% [vol/vol] Tween 20, and 5% [wt/vol] nonfat dried milk powder) prior to addition of primary antibody in the same blocking solution. Primary antibodies were detected with alkaline phosphatase-conjugated goat anti-rabbit immunoglobulin G (whole molecule; Sigma) treated with CDP-star chemiluminescent substrate (Amersham International) as directed by manufacturer.

RESULTS

Cloning of the R. meliloti phn genes.

The sequence of a putative R. meliloti phn gene cluster (28) was used to design PCR primers to amplify this region. Primers RmPhn1 and RmPhn2 (Fig. 1; Table 2) used in a PCR with R. meliloti genomic DNA gave an expected 3.9-kb product. Products from independent reactions were cloned to give plasmids pRMP1 and pRMP2, respectively. The inserts were restriction mapped to provide a comparison with the submitted sequence (Fig. 1). We further confirmed by Southern hybridization that the cloned DNA represents the genome correctly (data not shown). The DNA cloned into pRMP1 and pRMP2 was sequenced and found to be identical to that already submitted (28) between nucleotides (nt) 268 and 4148, except for two differences (T to C at nt 852 and G to A at nt 926). TestCode prediction and comparison with a codon usage table confirm that homologues of E. coli PhnG, PhnH, PhnI, PhnJ, and PhnK proteins are the most likely products encoded by this DNA (Table 3). However, we predict that the PhnG protein is 34 amino acid residues longer at the N terminus than that reported in GenBank (28). This additional sequence shows good homology to the E. coli PhnG protein.

TABLE 3.

Properties of predicted gene products in the R. meliloti phn gene cluster

Gene	Shine-Dalgarno sequencea	Predicted size of product		Closest homolog of predicted product	Properties
		Amino acids	Molecular mass (kDa)
phnG	GGAG (4)	155	16.6	E. coli PhnG protein; 37% identity and 77% similarity over 148-aab overlap
phnH	GGAGAGGA (3)	199	21.2	E. coli PhnH protein; 37% identity and 73% similarity in 188-aa overlap
phnI	GGAG (7)	367	40.8	E. coli PhnI protein; 57% identity and 88% similarity in 354-aa overlap	Contains a sequence with similarity to lipoxygenase pattern 2 sequence
phnJ	AGGAG (7)	297	33.4	E. coli PhnJ protein; 68% identity and 97% similarity in 275-aa overlap	Has four conserved cysteines, a potential metal-binding site
phnK	AGGAGG (7)	258	28.6	E. coli PhnK protein; 68% identity and 90% similarity in 253-aa overlap	HisP-like nucleotide-binding protein

^aNumbers in parentheses indicate nucleotide spacing from the putative Shine-Dalgarno sequence to the initiating methionine. 

^baa, amino acid. 

PhnI and PhnJ are especially well conserved between E. coli and R. meliloti (Table 3). It has been suggested that C-P bond cleavage by C-P lyase involves free radical- and redox-dependent chemistry, possibly involving transition metals (10). Metal ions could therefore be of considerable importance in C-P lyase, and therefore it is interesting that alignments of PhnI reveal that the C-terminal domain contains histidinyl residues which show an arrangement resembling that of a lipoxygenase (Fig. 2a) and which may provide ligands to metal ions. Also, alignments of PhnJ reveal four conserved cysteinyl residues, which could be ligands to a metal or metal sulfur cluster site (Fig. 2b).

FIG. 2.

(a) Alignment of E. coli PhnI (EcPhnI) and R. meliloti PhnI (RmPhnI) proteins. The highlighted amino acids are the additional C-terminal sequence on the R. meliloti protein and a region that shows homology to part of lipoxygenase (LOX) L-2 of Oryza sativa (SwissProt accession no. P29250) which contains the iron-binding region signature 2 (Prosite PS00081). (b) Alignment of E. coli and R. meliloti PhnJ proteins. The four conserved cysteines are arrowed, and the additional C-terminal sequence on the R. meliloti protein is highlighted.

Construction and characterization of a mutant in the putative phn operon.

To discover if the DNA cloned as described above is necessary for phosphonate utilization, we constructed a strain which has a 2-kb deletion within this region of the genome (Fig. 1). The presence of a genomic lesion was confirmed by Southern hybridization and by PCR amplification of genomic DNA from the mutant (data not shown). The phenotype of this mutation was assayed by comparing the rates of growth in liquid minimal medium of the mutant and wild type on several phosphonates and other compounds as sole sources of phosphorus (Table 4). This mutation renders R. meliloti unable to utilize aminomethylphosphonate, 3-aminopropylphosphonate, glyphosate, methylphosphonate, ethylphosphonate, propylphosphonate, phenylphosphonate, benzylphosphonate, 4-aminobenzylphosphonate, and PA. The mutant strain is, however, still able to use 2-AEP and glycerol-3-phosphate. This result demonstrates that the genomic lesion it carries specifically affects phosphonate metabolism but does not disrupt the PhoB-mediated phosphorus starvation response in this organism (2, 29).

TABLE 4.

Abilities of Rm1021 and Pn1 R. meliloti strains to grow on a various phosphorus sources compared with results of previous studies with E. coli

Phosphorus source	Growth ofa:
	Rm1021	Pn1	E. coli (reference[s])
Aminoethylphosphonate	+++b	+++b	G (23, 33)
Aminomethylphosphonate	+++b	−b
3-Aminopropylphosphonate	++	−
Glyphosate	++b	−b	NG (18)
Methylphosphonate	++b	−b	G (10)
Ethylphosphonate	++	−	G (10)
Propylphosphonate	++	−	G (42)
tert-butylphosphonate	−	−	NG (10)
Phenylphosphonate	++	−	G (10)
Benzylphosphonate	++	−	NG (10)
4-Aminobenzylphosphonate	++	−
Phosphonoacetate	+b	−b	G (33)
Glycerol-3-phosphate	+++	+++

^a+++, similar rate of growth and final OD as the positive control culture growing with 0.5 mM KH₂PO₄ as a phosphorus source in liquid minimal medium; ++, lower rate of growth than the positive control culture but reaching the same final OD; +, lower rate of growth than the positive control culture and lower final OD; −, similar rate of growth and final OD as the negative control containing no added phosphorus source; G, growth; NG, no growth. 

^bResult confirmed on minimal medium solidified with phosphate-free agarose. 

Overexpression of the putative phn genes of R. meliloti.

We made expression constructs designed to produce His-tagged and native forms of the putative Phn proteins (Table 1) and confirmed each structure by nucleotide sequencing (data not shown). Three of the five gene constructs, phnG, phnH, and phnK, produced large amounts of recombinant protein irrespective of the presence of the His tag (Fig. 3). Overexpression of putative PhnG, -H, and -K polypeptides lacking the N-terminal fusion resulted in the production of recombinant proteins of 16, 22, and 26.9 kDa, respectively, in agreement with predicted values (Table 3). By contrast, no visible overexpression was achieved with the other two gene constructs (phnI and phnJ [data not shown]). Many reasons exist for the lack of expression of recombinant proteins (41); however, the lack of expression of phnI and phnJ of R. meliloti may be significant. Possibly the whole C-P lyase complex is required to stabilize these components of the complex. Polyclonal antibodies were raised against the affinity-purified His-tagged recombinant proteins and used in subsequent experiments.

FIG. 3.

Expression of phn genes in E. coli BL21(DE3). Proteins were analyzed by SDS-PAGE as described in Materials and Methods. Lanes contain cell extracts of E. coli BL21(DE3) containing expression constructs induced with 1 mM IPTG. Lanes: 1, pTHGB (PhnG with His tag); 2, pTHGC (PhnG without His tag); 3, pTHHB (PhnH with His tag); 4, pTHHC (PhnH without His tag); 5, pTHKB (PhnK with His Tag); 6, pTHKC (PhnK without His tag). Molecular masses (in kilodaltons) are shown to the left; expressed proteins are indicated by arrows.

Expression of putative phn genes in R. meliloti.

Western immunoblots were made of R. meliloti cell extracts grown on liquid minimal media in the presence of various phosphorus sources. The results show that the antibodies to the three putative phn genes, phnG, -H and -K, do indeed detect polypeptides of the predicted size (Fig. 4). This finding suggests that the genes do not undergo any major posttranslational modification, e.g., removal of signal sequence. All three polyclonal antisera gave consistent results in that bands corresponding to those of the putative phn gene products were visualized only in cell extracts of R. meliloti grown in the presence of methanephosphonate, aminomethylphosphonate, and glyphosate as sole sources of phosphorus. Bands were absent from cell extracts of R. meliloti grown in the presence or absence of inorganic phosphate and aminoethylphosphonate. These data obtained from the Western blots are consistent with those for R. meliloti Pn1 (Table 4) with regard to growth and expression.

FIG. 4.

Immunoblot analysis of cell extracts from R. meliloti grown on various phosphorus sources. Lanes: 1, no phosphorus source; 2, inorganic phosphate; 3, methylphosphonate; 4, aminomethylphosphonate; 5, aminoethylphosphonate; 6, glyphosate. Lanes 7 to 9 contain cell extracts of E. coli BL21(DE3) containing pTHGC (phnG), pTHHC (phnH), and pTHKC (phnK) induced with 1 mM IPTG. Immunoblotting was done with antisera raised against PhnG (a), antisera raised against PhnH (b), and antisera raised against PhnK (c).

Extracts from strain Pn1, induced by phosphorus starvation in the presence of phosphonates, were also probed with the three polyclonal antisera. The mutant produces wild-type levels of PhnG, reduced levels of PhnH, and no detectable PhnK (data not shown). This finding demonstrates (i) that the mutation carried by Pn1 is polar, abolishing expression of genes 3′ to the insertion, which suggests that these genes are regulated by single promoter as predicted in the original sequence submission (28), and (ii) that PhnG is stable in the absence of at least phnI, phnJ, and phnK gene products.

DISCUSSION

In this study, we confirm that the genome of R. meliloti contains a locus homologous to part of the phn gene cluster of E. coli and that mutagenesis of this locus prevents R. meliloti from degrading a wide variety of, though not all, phosphonates. Antibodies raised to the putative PhnG, PhnH, and PhnK proteins from R. meliloti were used to show for the first time in any organism that these proteins are indeed expressed. Moreover, their expression in R. meliloti was seen only during growth with particular phosphonates as sole phosphorus sources. These observations are consistent with R. meliloti having a C-P lyase-type enzyme encoded by a phn gene cluster similar, in part at least, to that of E. coli. This view is substantiated by our observation that R. meliloti produces methane when growing with methylphosphonate as the sole phosphorus source (15) since the direct evolution of hydrocarbon products from phosphonates is a hallmark of C-P lyase activity (10, 42).

The C-P lyase enzyme of R. meliloti appears to have a broader substrate specificity than its E. coli counterpart because while R. meliloti can grow with benzylphosphonate and glyphosate as sole phosphorus sources, E. coli cannot. In the case of glyphosate, R. meliloti gives products consistent with degradation involving C-P lyase activity (25). Growth on a similar wide range of phosphonates has been observed in Agrobacterium radiobacter (42) and in Arthrobacter sp. strain GLP1 (18), where the broad specificity was proposed to result from the activity of two different C-P lyases. Our data suggest that R. meliloti has a single C-P lyase which is able to degrade a wide range of phosphonates.

This work also sheds light on the variety of inducible phosphonate-degrading enzymes that R. meliloti may express under phosphorus starvation. R. meliloti, like Enterobacter aerogenes (23), appears to have an enzyme or enzymes other than C-P lyase that degrades 2-AEP, namely, phosphonatase, because disruption of the phn gene cluster does not prevent growth on 2-AEP. Also, phnG, -H, and -K gene products were not expressed in R. meliloti 1021 growing with 2-AEP as a phosphorus source. Therefore, the enzyme(s) that degrades 2-AEP appears to be biochemically and genetically distinct from C-P lyase.

As regards the third class of C-P bond-cleaving enzyme known, PA hydrolase (27), we have shown that R. meliloti cannot induce this type of enzyme when using PA as its sole phosphorus source because even the relatively slow growth with PA is abolished when C-P lyase is inactivated by mutation.

A full understanding of C-P lyase function requires a better knowledge of the role of each of the phn gene products. However, we are unable to make any inferences about the function of individual genes from the phn mutation that we have constructed. Apart from the deletion that we have shown, the mutation carried by strain Pn1 is polar. The absence of a candidate terminator downstream of the R. meliloti phnK gene suggests the existence of more phn genes downstream, and preliminary sequencing indicates the presence of a gene homologous to E. coli phnL 3′ to phnK (15). Nonetheless, the availability of antibodies specific to several putative phn gene products is an important step toward further characterization of the C-P lyase enzyme. The data presented in this report provide the first evidence that proteins predicted to be part of the C-P lyase complex are expressed in organisms provided with phosphonates as sole phosphorus sources. These antibodies are useful tools for examining the hypotheses about the properties, localization, and activity of C-P lyase.