Characterisation of Campylobacter jejuni genes potentially involved in phosphonate degradation

Potential biological roles of the Campylobacter jejuni genes cj0641, cj0774c and cj1663 were investigated. The proteins encoded by these genes showed sequence similarities to the phosphonate utilisation PhnH, K and L gene products of Escherichia coli. The genes cj0641, cj0774c and cj1663 were amplified from the pathogenic C. jejuni strain 81116, sequenced, and cloned into pGEM-T Easy vectors. Recombinant plasmids were used to disrupt each one of the genes by inserting a kanamycin resistance (KmR) cassette employing site-directed mutagenesis or inverse PCR. Campylobacter jejuni 81116 isogenic mutants were generated by integration of the mutated genes into the genome of the wild-type strain. The C. jejuni mutants grew on primary isolation plates, but they could not be purified by subsequent passages owing to cell death. The mutant C. jejuni strains survived and proliferated in co-cultures with wild-type bacteria or in media in which wild-type C. jejuni had been previously grown. PCR analyses of mixed wild-type/mutant cultures served to verify the presence of the mutated gene in the genome of a fraction of the total bacterial population. The data suggested that each mutation inactivated a gene essential for survival. Rates of phosphonate catabolism in lysates of E. coli strain DH5α were determined using proton nuclear magnetic resonance spectroscopy. Whole-cell lysates of the wild-type degraded phosphonoacetate, phenylphosphonate and aminomethylphosphonate. Significant differences in the rates of phosphonate degradation were observed between lysates of wild-type E. coli, and of bacteria transformed with each one of the vectors carrying one of the C. jejuni genes, suggesting that these genes were involved in phosphonate catabolism.


Background
Phosphorus is an essential element for living organisms. In bacteria it has roles in biosynthesis and energy metabolism, as well as in the structure of biomolecules. Phosphonates (Phn) are a class of organophosphates structurally similar to phosphate esters, but characterised by a carbon-to-phosphorus (C-P) bond [1][2][3]. The C-P bond is highly stable compared to oxygen-to-phosphorus, nitrogen-to-phosphorus or sulphur-to-phosphorus bonds. The C-P bond is resistant to chemical hydrolysis, thermal decomposition, photolysis [4], and to the action of phosphatases [5,6]. During phosphate limitation stress, some bacteria can utilise phosphonates as an alternative source of phosphorus by breaking the C-P bonds of these compounds [7][8][9].
Genes encoding C-P lyase pathways have been found in bacteria of the genera Enterobacter, Escherichia, Klebsiella and Kluyvera [1,[11][12][13][14]; and phosphonatase pathways are found in the genera Enterobacter and Salmonella [11,15]. Recently, it was demonstrated that Campylobacter spp. are capable of degrading some phosphonate-bearing compounds although in silico analyses indicated that no genes orthologous to those encoding C-P bond-cleaving enzymes in other bacteria are present in the genome of Campylobacter jejuni [12]. Two different C-P bond cleavage activities were discovered in Campylobacter spp. The first one is associated with the cell wall and has the ability to catalyse phenylphosphonate and phenylphosphinate (PhePhi). The second cleavage activity is found in both the cell wall and cytosolic fractions and catalyses phosphonoalkyl carboxylates [12]. Phosphonate metabolism in Campylobacter does not require phosphate starvation conditions, unlike other bacteria in which the expression of Phn-degrading enzymes are under control of the Pho regulon [7,13,16].
The molecular basis of Phn catabolism by C. jejuni wasinvestigated in this study. In silico analyses were performed to identify C. jejuni genes encoding proteins with similarity to those involved in C-P bond hydrolysis in E. coli K12 [8]. Isogenic mutants of wild-type C. jejuni strain 81116 were constructed by disrupting with a kanamycin cassette three genes encoding proteins potentially implicated in phosphonate degradation, and the viability of the mutants was determined. Mutants of E. coli DH5α were generated by transforming the wild-type strain with plas-mids carrying each one of the C. jejuni genes putatively involved in Phn catabolism. The rates of degradation of three phosphonate compounds by the mutant E. coli were compared to those of the wild-type.

In silico analyses of genes putatively involved in phosphonate catabolism
Three genes orthologous to those involved in C-P bond cleavage in E. coli were identified in the genome of C. jejuni strain NCTC 11168 by protein homology analyses. The proteins encoded by the C. jejuni genes, cj0774c and cj1663 showed homology to the E. coli membrane-associated C-P lyase proteins PhnK and PhnL encoded by the phnC-P operon (Table 1). Domain architecture searches demonstrated that the four proteins contained the ABC_MJ0796 domain characteristic of some ATP-binding cassettes of ABC transport systems. The protein encoded by cj0641 showed low similarity to E. coli PhnH, but had 48% similarity to the inorganic polyphosphate/ATP-NAD kinase JHP1433 of H. pylori strain J99 an enzyme which utilizes ATP and other nucleoside triphosphates as well as inorganic polyphosphate as a source of phosphorus, and contains a COG3221.2 domain of the periplasmic component of ABC-type/phosphonate transport system. In comparison, the similarities of E. coli PhnH, PhnK and PhnL with proteins of the C-P lyase pathway of Enterobacter sp. 638 were 92, 98 and 92%, respectively; and the similarities with Klebsiella pneumoniae proteins were 90, 97 and 90%, respectively.
Investigation of the genomic organisation of the genes cj0641, cj0774c and cj1663 indicated that they were not co-located, and none of the genes appeared to be co-transcribed with other genes.

Isolation of Campylobacter jejuni genes
The three coding regions cj0641, cj0774c and cj1663 were amplified from C. jejuni strain 81116 and cloned into the vector pGEM-T Easy by standard cloning techniques. The primers used for the amplification of these genes given in Table 2 were designed based on the published nucleotide sequence of C. jejuni NCTC 11168 [17]. Sequence analyses of the genes cloned into the recombinant plasmids pGU0303 (pGEM-T EasyΩcj0641), pGU0304 (pGEM-T EasyΩcj0774c) and pGU0305 (pGEM-T EasyΩcj1663) The recombinant plasmids with disrupted C. jejuni genes pGU0306 (pGEM-T EasyΩcj0641ΩKm R ), pGU0307 (pGEM-T EasyΩcj0774cΩKm R ) and pGU0308 (pGEM-T EasyΩcj1663ΩKm R ) were subsequently employed to generate isogenic mutants of C. jejuni 81116.

Generation and identification of C. jejuni 81116 isogenic mutants
The disrupted cj0641, cj0774c and cj1663 genes were introduced into the genome of wild-type C. jejuni by electroporating electro-competent C. jejuni 81116 cells with plasmid DNA containing the mutated genes. Transformed bacteria could only be recovered on primary isolation plates.
To verify the presence of isogenic mutants, the transformants were initially grown on a TSA/HBA recovery plates without antibiotic selection. Subsequent plating onto kanamycin selective media resulted in cell death. All attempts to recover pure cultures of the isogenic mutants were unsuccessful. PCR analyses of bacterial DNA from single colonies on the initial recovery plates indicated that the genes disrupted by the insertion of the Km R cassette were present in the genome of some of the cells, showing that the cultures which grew on the plates were a mixed bacterial cell population. Verification of the insertion of the kanamycin cassette was determined by amplification of the Km R cassette. Amplicons of either mutant or wildtype gene could not be recovered using the gene-specific primers from mixed populations ( Figure 1).
Since none of the isogenic mutants could be isolated but grew in mixed cultures on primary recovery plates, it was hypothesised that the mutants may require a product or products produced by the wild-type strain to survive. Mixed cultures of wild-type and mutant bacteria harvested from primary isolation plates were grown in liquid "conditioned media" with or without 50 μg/ml kanamycin. In controls, mixed bacterial cultures were grown in liquid fresh media with or without kanamycin. In the presence of kanamycin C. jejuni 81116 cells grew only in conditioned media ( Table 3). The C. jejuni bacteria isolated from cultures grown in conditioned media with kanamycin were plated onto TSA/HBA plates with or without 50 μg/ml kanamycin, but no bacteria were recovered from any these cultures, suggesting that bacteria without the kanamycin cassette were screened out by the conditioned media with kanamycin.

Measurement of phosphonate degradation by E. coli lysates
Degradation of PhePhn, phosphonoacetate (PhnAce), AmePhn and PhePhi, was measured in whole-cell lysates of wild-type E. coli DH5α grown on LBA plates. The catabolism of phosphonate was confirmed as resulting from enzyme activities by heating the lysates for 4 h at 80°C, or by suspending them in 1% SDS and observing the abolition of catabolic activity.
Degradation of PhePhn, PhnAce and AmePhn was measured in E. coli whole-cell lysates of bacteria transformed with either of the plasmids pGEM-T Easy pGU0303, pGU0304, pGU0305, pGU0306, pGU0307 or pGU0308  Figure 2). Similarly to untransformed wild-type bacteria, the lysates of cells transformed with pGU0303, pGU0304 and pGU0305 showed degradation of PhePhn and PhnAce, with significant reductions in the rates. AmePhn degradation was similar in cells transformed with pGU0303, pGU0304 and pGU0305 relative to the rates measured in the control. Transformation of E. coli with pGU0306, pGU0307 or pGU0308 had no effect on any of the three activities. In addition, no change was observed in the degradation of PhePhi between all strains ( Figure 2).

Discussion
Previous studies demonstrated that Campylobacter spp. catabolise various phoshonates [12], but sequence analyses of the C. jejuni strains 11168 and RM1221 genomes did not identify in this species orthologues of genes involved in phosphonate degradation in E. coli or Salmo-nella typhimurium. Nonetheless, the coding sequences cj0641, cj0774c and cj1663 showed some sequence or domain homology to genes encoding proteins of the C-P lyase of E. coli. These genes were disrupted in C. jejuni strain 81116 to determine whether the resultant mutants were altered in their ability to degrade phosphonates. The low sequence similarity between E. coli PhnH, PhnK and PhnL and C. jejuni proteins was in contrast with the high sequence similarity found with Enterobacter sp. 638 and K. pneumoniae proteins of the C-P lyase pathway, suggesting that the C. jejuni enzymes belong to different pathways.
The three genes from C. jejuni strain 81116 were cloned, and analyses of their sequences showed nucleotide identities of 96-99% with the cj0641, cj0774c and cj1663 genes of C. jejuni strain 11168. The data indicated that the genes were highly conserved, even more than the highly conserved flaA genes of Campylobacter spp. which have simi-Campylobacter jejuni single colonies from the initial isolation plate Figure 1 Campylobacter jejuni single colonies from the initial isolation plate. Lane 1: 1 kb ladder. Lane 2: pGU0303 amplified cj0641 using gene-specific primers. Lane 3: pGU0306 amplified Km R using Km R -specific primers. Lane 4: C. jejuni from the original isolation plate, amplified cj0641 using gene-specific primers. Lane 5: C. jejuni from the original isolation plate, amplified Km R using Km R specific primers. Lane 6-9: C. jejuni from the original isolation plate, using a combination of gene-and Km R -specific primers. Lane 10: Km R cassette. Lane 3: pGU0306 amplified Km R using Km R -specific primers.
Lane 5: C. jejuni mixed culture, amplified Km R using Km R specific primers.
Lane 6-9: C. jejuni mixed culture, using a combination of gene-and Km R -specific primers.
Lane 10: Km R casette.  larities between 73.6 and 92.3% [19]. The three genes were not co-located or co-transcribed in the C. jejuni 11168 or 81116 genomes, and did not form part of operons.
The genes cj0641, cj0774c or cj1663 were inactivated by inserting a Km R cassette, and C. jejuni 81116 isogenic mutants were constructed by incorporation of each the disrupted genes into the genome of the wild-type strain through double allelic exchanges. The mutant bacteria were not viable, but were able to grow in mixed cultures with the wild-type strain. Growth experiments suggested that the mutants required for viability some factor(s) expressed by the wild-type strain. It is highly unlikely that a mutation in each of the three genes would result in identical phenotypes by chance. Inactivation of these genes by insertion of a Km R cassette may have resulted in polar mutations. However, none of three genes belongs to a recognisable cluster or to an operon, and analyses of the genes downstream of cj0641, cj0774c and cj1663 indicated that they were not co-transcribed. In addition, the insertion of the Km R cassette has been performed in numerous studies without having polar effects on downstream genes [18,20]. Thus, polar mutations probably were not the cause of the decreased survival. Alternatively, it is possible that these genes encode multifunctional proteins given that the C. jejuni genome is one of the most densely transcribed genomes known to date [17], and some of the functions of those proteins could be essential to the survival of C. jejuni.
The results of the bacterial cell culture experiments indicated the dependence of mutated C. jejuni on the wildtype strain for survival, suggesting that the mutants lacked the ability to produce, regulate, or utilise some compound(s) produced by the wild-type C. jejuni strain 81116.
Indirect evidence for the involvement of the proteins CJ0641, CJ0774c and CJ1663 in phosphonate degradation was obtained by expressing them in E. coli DH5α transformed with plasmids carrying each one of the genes. The C-P lyase enzyme of E. coli is expressed under conditions of phosphate starvation [6]. To ensure that this endogenous E. coli lyase was not expressed, bacteria were grown in the presence of inorganic phosphate. Hence, the catabolism of phosphonates observed in the transformed E. coli was the result of other enzyme activities ( Figure 2).
Transformation of the wild-type E. coli strain with the plasmid pGEM-T Easy did not affect the rates of degradation of PhePhn, PhnAce and AmePhn. Transformation with a plasmid bearing cj0641, cj0774c or cj1663 altered the rate of PhePhn and PhnAce degradation relative to the wild-type (Figure 2). Transformation with a plasmid bearing one of the three genes inactivated with a kanamycin cassette had no effect on phosphonate degradation (Figure 2). The largest inhibition of PhePhn degradation was observed in E. coli transformed with cj0641, while the largest inhibition of PhnAce was due to transformation with cj1663. The relative decrease of phosphonate degradation between the three strains suggested that CJ0641 was more specific to PhePhn and CJ1663 was more specific to PhnAce. Comparison of the relative decrease of phosphonate degradation between substrates for each strain showed that E. coli transformed with pGU0303 and pGU0304 decreased PhnAce and PhePhn catabolism in equal amounts, while pGU0305 inhibited PhnAce degradation significantly more than that of PhePhn.
A reasonable expectation for the expression of an exogenous enzyme in a bacterium which also has a similar endogenous activity is that the overall activity measured in the transformed bacterium will be higher than in the parent organism. However, the measured activity is modulated by a number of factors including the relative abundance of both enzymes, their affinity for the substrate, their velocities and the substrate concentration. Regarding the latter, the overall measured activity will be higher when the concentration of the substrate is sufficient for both the endogenous and exogenous enzymes to operate as if the other were not present. Otherwise, both enzymes will compete for the substrate and depending on their relative affinities for the substrate and velocities, the overall measured activity may be higher, not significantly changed or lower. The latter case could occur when the exogenous enzyme has a much greater affinity and significantly lower velocity. In the case of PhePhn and PhnAce the substrate concentrations are well below saturation of the C. jejuni enzymes [12], hence the enzymes expressed in E. coli will compete with the endogenous activities, and the observed decrease of overall rates in the transformed bacteria could be explained by a competition of enzymes for the substrate Several factors could be responsible for the lack of change in the rates of AmePhn degradation. C. jejuni has an enzyme system that degrades PhePhn and PhnAce and an independent system that degrades AmePhn, with the catabolism of AmePhn approximately half of that of the other two phosphonates [12]. This suggested that the former activities were due to different enzyme(s) than that of the latter. Also, the absence of changes in the rates of AmePhn degradation may be due to homologues of the broad specificity C-P lyase of E. coli [8] having different specificities in C. jejuni. Finally, the rate of AmePhn degradation may have undergone changes which were below the levels of detection owing to its significantly smaller value than the rate of PhePhn and PhnAce.
No change was observed in the degradation of PhePhi between all strains showing that the changes in activity are specific to phosphonate degradation, thus, supporting the interpretation that these genes are involved in phosphonate metabolism.

Conclusion
The data showed that the gene products of cj0641, cj0774c and cj1663 were essential for the survival of the bacterium. In addition, evidence supported that CJ0641, CJ0774c and CJ1663 would be involved in phosphonate metabolism in the bacterium. The specific molecular events leading to changes in phosphonate degradation in transformed E. coli remain to be fully elucidated, but the study demonstrated that the C. jejuni proteins expressed in E. coli modulated the rates of phosphonate catabolism in the latter.

Bacterial strains, plasmids and growth conditions
The C. jejuni and E. coli strains, and the plasmids [21] used in this study are listed in Table 4. Campylobacter jejuni strains NCTC 11168 [22] and 81116 [23] were grown on Columbia Agar supplemented with 5% (v/v) defibrinated horse blood (HBA), or TSA supplemented with 5% (v/v) defibrinated horse blood and ampicillin or kanamycin when required. C. jejuni cells were incubated for 48 hours at 37°C under the microaerobic conditions 90% N 2 , 5% CO 2 , and 5% O 2 . Liquid cultures of C. jejuni were grown in Brucella Broth, at 37°C for 24-48 hours under the same microaerobic conditions as for plates. Cell growth was determined by measuring optical densities (OD) at 410 nm.
"Conditioned" Brucella Broth to grow mutant C. jejuni bacteria was prepared by growing wild-type C. jejuni strain 81116 in Brucella Broth for 24 hours, removing the cells by centrifugation, and collecting and filtering the supernatant. The absence of viable cells in the supernatant was verified by plating it and observing no growth on the plates.
Escherichia coli strain DH5α was grown on Luria Broth Agar (LBA) plates supplemented with ampicillin, kanamycin, IPTG or X-Gal when required, and incubated at 37°C for 24 hours. Liquid cultures of E. coli were grown in Luria Broth with antibiotic supplements, ampicillin or kanamycin, where necessary, and under the same conditions as for plates.

Genome searches
Sequences of E. coli K12 proteins [24] were employed to search in the fully annotated genomes of C. jejuni strains NCTC 11168 and RM1221, as well as the genomes of Klebsiella pneumoniae and Enterobacter sp. 638 for genes orthologous to those involved in C-P bond hydrolysis. The E. coli sequences were those encoded by the genes phnG, phnH, phnI, phnJ, phnK, phnL, and phnM (accession numbers P16685, P16686, P16687, P16688, P16678, P16679, and P16689) that are required for alkylphosphonate catalysis and likely constituting a membrane-associated carbon-phosphorus (C-P) lyase [2,7].

PCR amplifications
Primers for the genes cj0641, cj0774c and cj1663 in C. jejuni strain 81116, were designed using the published nucleotide sequence of C. jejuni strain NCTC 11168 [17]. Primer sequences are shown in Table 2. Protein similarity analyses comparing the sequenced genes of C. jejuni 81116 to the published sequences of C. jejuni 11168 were performed using BLASTP program for microbial genomes of the NCBI.
Standard PCR in 50 μl reaction mixtures were performed with the following parameters: 95°C for 4 minutes, 30 cycles of 94°C for 30 s, 52°C for 30 s, and 72°C for 2 min. Template DNA was prepared by the crude lysis boiling method, in which a single C. jejuni colony was boiled for 5 min in 150 μl of water and cellular debris were removed by centrifugation.

Cloning and screening of recombinant plasmids pGU0303, pGU0304 and pGU0305
Amplified PCR fragments were cloned into the bacterial cloning vector system pGEM-T Easy (Promega; Madison, WI, USA) in E. coli by standard cloning techniques [25]. The plasmids pGU0303, pGU0304 and pGU0305 each containing one of the three C. jejuni genes are described in Table 4. Transformant colonies putatively carrying recombinant plasmids were screened by blue/white colony selection, and further verified by PCR, using plasmid specific primers for the T7 and SP6 polymerase promoters, which flank the multiple cloning region of pGEM-T Easy. Prior to mutagenesis, constructs were sequenced by didioxynucleotide sequencing, to verify the correct base pair order.
In vitro mutagenesis of cloned Campylobacter jejuni genes cj0641, cj0774c and cj1663 Mutagenesis was performed by insertional inactivation of the genes cj0641, cj0774c and cj1663 using a non-polar kanamycin antibiotic resistance cassette Km R [26,27]. The Km R cassette was amplified from the plasmid pMW2 using primers with BglII sites incorporated into the termini ( Table 2). The Km R cassette was inserted into unique restriction sites within the cloned C. jejuni inserts. Linearised plasmids for pGU0303 (pGEM-T EasyΩcj0641), and pGU0304 (pGEM-T EasyΩcj0774c) were generated by restriction enzyme digestion with ClaI and HindIII respectively, according to the manufacturer's instructions.
The plasmid pGU0305 (pGEM-T EasyΩcj1663) was linearised by inverse PCR, as no unique restriction sites existed within cj1663. The primers cj1663p2A INV and cj1663p2B INV were designed based on the nucleotide sequence of pGU0305 and created a small deletion within cj1663. Inverse PCR in 50 μL reaction mixtures was conducted as follows: 95°C for 5 min, 40 cycles 94°C for 30 s, 52°C for 1 min, and 72°C for 5 min. The Eppendorf Tri-pleMaster Taq was the DNA polymerase used for inverse PCR (Eppendorf; North Ryde, NSW, Australia).

Campylobacter jejuni isogenic mutants
Electrocompetent C. jejuni strain 81116 cells were prepared [28] and transformed with mutated recombinant plasmids using a BioRad MicroPulser (BioRad; Regents Park, NSW, Australia) at a voltage of 2.48 V, and time constant of 5 min. The contents of the 2 mm cuvettes were flushed with BHI broth and allowed to recover at 37°C for 5 h under microaerobic conditions. Reactions were plated onto TSA/HBA plates.

Measurement of phosphonate catabolism using nuclear magnetic resonance spectroscopy
Wild-type and transformed E. coli strain DH5α cells were grown on LBA plates supplemented with 50 μg/ml ampicillin and incubated at 37°C for 6 h. The bacteria were harvested in 150 mM sodium chloride (NaCl) and centrifuged 16,000 × g at 4°C for 10 min. The pellet was collected and the supernatant discarded. The pellet was resuspended in 14 ml 150 mM NaCl solution and washed twice more. Cells were lysed by twice freezing in liquid nitrogen and thawing.
Catabolism of AmePhn, PhePhn, PhePhi and PhnAce was measured employing proton nuclear magnetic resonance ( 1 H-NMR) spectroscopy. Lysates (150 μl) were suspended in 150 mM NaCl and mixed with 2 H 2 O (50 μl), 150 mM KCl (50 μl), and 150 mM NaCl (150 μl) before phosphonate was added at time zero to give a final volume of 600 μl. Initial concentrations of phosphonate substrates were 120 mM. Suspensions of bacterial lysates were placed in 5 mm NMR tubes (Wilmad; Buena, NJ, USA) and measurements of enzyme activities were carried out at 37°C. 1 H-NMR free induction decays were collected using a Bruker DMX-600 spectrometer, operating in the pulsed Fourier transform mode with quadrature detection. The instrumental parameters for the spectrometer were: operating frequency 600.13 MHz, spectral width 6009.61 Hz, memory size 16 K, acquisition time 3.61 s, number of transients 64, pulse angle 50° (ca. 3 μs) and relaxation delay with solvent presaturation 1.7 s. Spectral resolution was enhanced by Gaussian multiplication with line broadening of 0.7 Hz and Gaussian broadening factor of 0.19. Proton spectra were acquired with presaturation of the water resonance. The time-evolution of substrates and products were followed by acquiring sequential spectra of the reactions. Enzyme rates were obtained by measuring the decrease over time of the intensity of resonances aris-