Analysis of genes involved in methyl halide degradation in Aminobacter lissarensis CC495

Abstract

Aminobacter lissarensis CC495 is an aerobic facultative methylotroph capable of growth on glucose, glycerol, pyruvate and methylamine as well as the methyl halides methyl chloride and methyl bromide. Previously, cells grown on methyl chloride have been shown to express two polypeptides with apparent molecular masses of 67 and 29 kDa. The 67 kDa protein was purified and identified as a halomethane:bisulfide/halide ion methyltransferase. This study describes a single gene cluster in A. lissarensis CC495 containing the methyl halide utilisation genes cmuB, cmuA, cmuC, orf 188, paaE and hutI. The genes correspond to the same order and have a high similarity to a gene cluster found in Aminobacter ciceronei IMB-1 and Hyphomicrobium chloromethanicum strain CM2 indicating that genes encoding methyl halide degradation are highly conserved in these strains.

Methyl chloride, Methyl bromide, Aminobacter

1 Introduction

The methyl halides, methyl chloride (MeCl) and methyl bromide (MeBr), are trace gases in the troposphere which, on transport to the stratosphere, undergo photolysis to release halogen atoms which catalytically destroy ozone [1]. An important sink in the environment for methyl halides is biological degradation, and bacteria have been isolated that use methyl halides as a sole source of carbon and energy. These include the terrestrial strains Methylobacterium chloromethanicum CM4 and Hyphomicrobium chloromethanicum CM2 [2,3], Aminobacter ciceronei IMB-1 [4–6], Aminobacter lissarensis CC495 [6–8] and the marine strain Leisingera methylohalidivorans [9,10].

A. lissarensis CC495 was isolated from the upper 5 cm of the soil horizon in a wood in County Down, Northern Ireland [7] and is closely related to A. ciceronei IMB-1, a methyl halide degrader isolated from MeBr fumigated agricultural soil [5]. The physiology and biochemistry of methyl halide degradation in A. lissarensis CC495 has been studied [7,8]. As in other methyl halide degrading bacteria, growth on methyl chloride was inducible [7]. Cells grown on methyl chloride expressed two polypeptides with apparent molecular masses of 67 and 29 kDa. The 67 kDa protein was purified and characterised as a halomethane:bisulfide/halide ion methyltransferase [7]. The enzyme is a corrinoid protein and the N-terminal sequence showed identity to the N-terminal sequence of CmuA from M. chloromethanicum CM4, H. chloromethanicum CM2 and the derived N-terminal sequence from A. ciceronei IMB-1.

The other three methyl halide degrading terrestrial strains have been studied at the physiological and molecular level and were found to possess the methyl halide utilisation gene cmuA as well as other genes involved in methyl halide utilisation. Biochemical and genetic experiments on M. chloromethanicum CM4 [11] have elucidated a pathway for methyl halide utilisation in this organism where the N-terminal of CmuA acts as a methyltransferase I enzyme and transfers the methyl group from methyl chloride to a corrinoid cofactor which is attached to the C-terminal of CmuA. CmuB then acts as a methyltransferase II enzyme and transfers the methyl group to tetrahydrofolate (H₄F). This folate-linked methyl group is then progressively oxidised to formate and finally to CO₂. Interestingly, the corrinoid protein derived from cmuA in M. chloromethanicum CM4 also exhibited the methyltransferase activity characteristic of the halomethane:bisulfide/halide ion methyltransferase from A. lissarensis CC495 [12].

Mutagenesis studies have shown that cmuB, cmuC, cmuA and purU are all essential for methyl chloride degradation in M. chloromethanicum CM4. Recent work on H. chloromethanicum CM2 using transposon mutagenesis, marker exchange mutagenesis and reverse transcription (RT)-PCR analysis [13] has shown that it metabolises MeCl using the same corrinoid specific pathway as M. chloromethanicum CM4 and that cmuB, cmuA, cmuC, fmdB, paaE, hutI and metF are co-transcribed and co-regulated. The aim of this study was to determine if the gene encoding the corrinoid halomethane:bisulfide/halide ion methyltransferase in A. lissarensis CC495 was similar to cmuA encoding the corrinoid H₄F methyltransferase I and to ascertain if strain CC495 had other genes encoding methyl halide utilisation in common with these strains.

2 Materials and methods

2.1 DNA extraction and PCR amplification

DNA extraction of chromosomal DNA from A. lissarensis CC495 was achieved using a method previously described for methanotrophs [14]. Plasmid DNA from recombinant Escherichia coli cells was extracted using the QIAGEN Midi kit (QIAGEN Ltd., Halden, Germany) according to the manufacturer's instructions. PCR primers to amplify cmuA in A. lissarensis CC495 were designed using cmuA sequences from M. chloromethanicum CM4, H. chloromethanicum CM2 and A. ciceronei IMB-1. These were cmuA802f (5′-TTC AAC GGC GAY ATG TAT CCY GG-3′) and cmuA1609r (5′-TCT CGA TGA ACT GCT CRG GCT-3′) [15]. PCR amplifications were performed in 50 μl (total volume) mixtures in 0.5 ml microcentrifuge tubes using a Hybaid touchdown thermal cycling system.

2.2 Construction of partial libraries of DNA from A. lissarensis strain CC495

Partial genomic libraries were constructed by cloning DNA from A. lissarensis CC495 into the multicopy vector pUC18. DNA from A. lissarensis CC495 was digested with various restriction enzymes and fragments of the required size were excised from gels, ligated into the cloning vector and transformed into E. coli. Fragments suitable for cloning were identified by Southern blotting as described [16] using radioactively labelled cmuA probes from A. ciceronei IMB-1. Colonies containing the cmuA gene were identified by colony blotting as described previously [16].

2.3 Sequencing and analysis

DNA sequencing was performed by cycle sequencing with a Dye Terminator kit (PE Applied Biosystems, Warrington, UK) and DNA sequences were analysed with a model 373A automated sequencing system (PE Applied Biosystems). DNA sequences and derived amino acid sequences were analysed using the DNAstar package. Similarity searches were performed using the gapped Basic Local Alignment Search Tool (BLAST) program [17] against public protein and gene databases. Alignments of DNA and amino acid sequences were made using the CLUSTAL W program. The sequence for the gene cluster from A. lissarensis CC495 has been deposited in GenBank accession number AY838881.

3 Results and discussion

3.1 Cloning and sequencing of methyl halide utilisation genes in A. lissarensis CC495

DNA extracted from A. lissarensis CC495 was used to identify cmuA-like genes using PCR primers designed from cmuA from M. chloromethanicum CM4, H. chloromethanicum CM2 and A. ciceronei IMB-1. These primers, cmuA802f and cmuA1609r, amplified a product of the correct size (807 bp) from strain CC495 which when sequenced, exhibited 83% identity to cmuA from A. ciceronei IMB-1. Southern hybridisation of A. lissarensis CC495 genomic DNA with a cmuA probe from A. ciceronei IMB-1 revealed four potential restriction fragments. An Eco RI (2.0 kb) fragment and a Pst I (6.4 kb) fragment were targeted for cloning and were ligated into pUC18 and transformed into E. coli TOPO10 cells to create partial clone libraries. The libraries were probed with the cmuA probe, identifying two clones (pCC2 and pCC6) which were sequenced (Fig. 1).

1

The methyl halide utilisation gene clusters from A. lissarensis CC495, A. ciceronei IMB-1, H. chloromethanicum CM2 and M. chloromethanicum CM4. CM4 I and CM4 II are the two clusters of genes cloned from M. chloromethanicum CM4 which contained genes involved in methyl halide utilisation.

The 6.4 kb gene cluster contained several open reading frames (Fig. 1). Six open reading frames corresponded largely with those found in the gene cluster in A. ciceronei IMB-1 (Table 1). Putative Shine–Dalgarno ribosomal binding sites were identified by visual inspection for cmuC, cmuA and paaE. Bearing in mind the biochemical evidence for a halomethane:bisulfide/halide ion methyltransferase in strain CC495, it is surprising that the gene order is identical to that found in A. ciceronei IMB-1 and H. chloromethanicum CM2 where the H₄F methylation route has been postulated.

1

Open reading frames within the methyl halide gene cluster ofAminobacter lissarensis strain CC495

Gene/orf	Length (aa)	Gene start–end (bp position)	Inferred function	Sequence comparison of representative protein hit (% identity)
cmuB	>294	Start–884	Methyltransferase	M. chloromethanicum CM4 CmuB (65%) over 294 aa
cmuC	413	881–2122	Methyltransferase	Partial sequence from A. ciceronei IMB-1 CmuC (56%), H. chloromethanicum CmuC (37%)
cmuA	619	2143–4003	Methyltransferase/corrinoid	A. ciceronei IMB-1 CmuA (83%)
orf188	188	3821–4387	Unknown	A. ciceronei IMB-1 Orf 146 (66%)
paaE	364	4405–5499	Reductase	A. ciceronei IMB-1 PaaE (78%)
hutI	>246	5675–end	Imidazolonepropionase	A. ciceronei IMB-1 HutI (61%) over 246 aa

3.2 CmuB

The region sequenced (see Fig. 1) corresponded to a translated product of 294 amino acids and showed highest identity to CmuB from M. chloromethanicum CM4 (65% over 294 aa), with 61% identity to CmuB from H. chloromethanicum CM2 [18]. CmuB in these organisms is 311 amino acids in length, which suggested that the majority of the gene in strain CC495 had been cloned. In M. chloromethanicum CM4, CmuB has been purified under anoxic and reducing conditions and identified as a methylcobalamin:H₄ folate methyltransferase II [11,19]. Besides CmuB from other methyl halide utilisers, the next most similar protein was MtrH from Methanobacterium thermoautotrophicum [20]. MtrH is a subunit of the methyl tetrahydromethanopterin (H₄MPT):coenzyme M complex and is thought to catalyse the transfer of a methyl group from H₄MPT to a cognate corrinoid binding protein [21]. This is analogous to the reverse of the reaction catalysed in methyl halide degradation by CmuB in M. chloromethanicum CM4; although the latter involves H₄F rather that H₄MPT.

3.3 CmuC

No ribosome binding site could be identified 5′ of cmuC. A putative start codon was found at base 881 making the polypeptide 413 aa in length. This made it longer than CmuC in H. chloromethanicum CM2 (370 aa) and M. chloromethanicum CM4 (378 aa) although the same length as Orf 414 located in M. chloromethanicum CM4 with which it had 30% identity. CmuC in strain CC495 had 56%, 37% and 28% identity to CmuC in A. ciceronei IMB-1, H. chloromethanicum CM2 and M. chloromethanicum CM4, respectively [11,13]. The function of CmuC in the other methyl halide utilisers is unknown although it is essential for growth on methyl chloride in M. chloromethanicum CM4 and in this strain was found to have identity with the methyltransferase MtbA from Methanosarcina barkerii [11].

3.4 CmuA

CmuA in strain CC495 had 83%, 78% and 77% identity to CmuA from A. ciceronei IMB-1, H. chloromethanicum CM2 and M. chloromethanicum CM4, respectively. It was 619 aa in length and as with the other genes appeared to be bifunctional with the N-terminal region acting as a methyltransferase and the C-terminal region being a corrinoid binding protein. The highest hit in a BLAST search from non-methyl halide utilising bacteria was a hypothetical multidomain protein from Methanosarcina acetivorans whose N-terminal domain was a methyltransferase and whose C-terminal domain was a corrinoid binding protein (NC003552). As CmuA in A. lissarensis CC495 shows very high identity to the CmuA sequences from the other methyl halide utilisers, consequently the N-terminal of the sequence has similarities to methyltransferases from methanogenic archaea whilst the C-terminal sequence shows similarity to corrinoid binding proteins. The methyltransferase region of the sequence is aligned with MtbA (CmtM) and MtsA. MtbA is one of two isoenzymes (MtaA and MtbA) found in M. barkeri which act as methyltransferases catalysing the transfer of a methyl group from a methylated corrinoid protein to coenzyme M [22]. MtsA is also found in M. barkeri and is homologous to MtaA and MtbA, also acting as a methyltransferase in the methylation of coenzyme M by methylated corrinoid protein [23]. All three enzymes are zinc-containing metalloproteins which contain one mole of zinc per mole of protein and possess the putative zinc binding motif (H–X–CX_n–C) [24,25]. This motif is also seen in CmuA from M. chloromethanicum CM4 [26] suggesting that CmuA may also be able to bind zinc and indeed the purified CmuA protein has been shown to contain 0.9 mol zinc per mole of protein [12]. Significantly, no evidence for the presence of zinc was found in purified CmuA from A. lissarensis CC495 [7]. The function of zinc in MtaA is believed to be to lower the pK_a of the coenzyme M thiol group that acts as methyl acceptor [27]. The relatively low pK_a of H₂S may render zinc unnecessary with HS⁻ acting as methyl acceptor with the enzyme from A. lissarensis CC495. The 67 kDa polypeptide induced when CC495 was grown on methyl chloride was N-terminal sequenced [7] and is identical to the N-terminal region of the derived amino acid sequence of cmuA sequenced in this study.

3.5 Orf 188

As with cmuC, a putative ribosome binding site for orf188 could not be identified. The putative start codon for this open reading frame is at bp 3821. Highest sequence identities of the derived polypeptide were with the putative transporter from A. ciceronei IMB-1 orf146 (66% identity) [28] and the transcriptional regulator fmdB from H. chloromethanicum CM2 (44% identity) [18].

3.6 PaaE

paaE encodes a 364 aa polypeptide with 78% and 53% identity to PaaE from A. ciceronei strain IMB-1 [28] and H. chloromethanicum CM2 [18]. Significant identity (30%) was also seen with a putative phenylacetic acid degradation NADH oxidoreductase from Streptomyces coelicolor (CAC44653) and with PaaE from E. coli (X97452) and Bradyrhizobium japonicum (NP769535) (Fig. 2). Conserved regions are marked which are similar to those found in the reductase components of the class 1A dioxygenase family of proteins and include an FMN/FAD binding site, an NAD binding domain and a [2Fe–2S] binding domain, although it seems that the genes encoding PaaE have conserved regions which do not match the ones of importance from the dioxygenase family, suggesting that there are other important residues specific to the PaaE proteins. The close proximity of paaE to genes involved in methyl halide degradation suggests that PaaE might also be involved. Recently paaE from H. chloromethanicum CM2 has been shown to be part of a cmuBCA metF operon that is co-transcribed and co-regulated [13].

2

Alignment of the putative PaaE sequence from A. lissarensis CC495 with the putative PaaE sequence from A. ciceronei IMB-1, the partial PaaE sequence from H. chloromethanicum CM2 and PaaE sequences from Escherichia coli (X97452), Bradyrhizobium japonicum (NP769535) and Streptomyces coelicolor (CAC44653). The FMN/FAD, NAD and [2Fe–2S] ferrodoxin conserved binding sites are indicated underneath the sequence [29,30]. Residues identical or similar in all six sequences are black or shaded, respectively.

3.7 HutI

The complete sequence for this open reading frame was not obtained since the 3′ end of this gene was not cloned. The sequence obtained encoded 246 aa and showed greatest identity (61%) to the putative imidazolonepropionase HutI from A. ciceronei IMB-1 [28] and has significant identity (25%) to HutI from Salmonella enterica subsp. Typhi (AL513382), Vibrio cholerae (NC002505), Yersinia pestis (AL590842) and Pseudomonas aeroginosa (AE004091).

Until recently the pathway for methyl halide utilisation via H₄F had been demonstrated only in M. chloromethanicum CM4. However, recent work by Borodina et al. [13] has provided strong genetic evidence that methyl halides are likely to be metabolised in a similar way in H. chloromethanicum CM2. The genetic data presented here provide strong evidence that methyl halide metabolism in A. lissarensis CC495 is likely to proceed by a similar enzyme mechanism to that found in A. ciceronei IMB-1, H. chloromethanicum CM2 and M. chloromethanicum CM4. However, the apparent paradox that CmuA in M. chloromethanicum CM4 mediates transfer of the methyl group to H₄F under anoxic and reducing conditions [19] yet CmuA from A. lissarensis CC495 (and indeed from M. chloromethanicum CM4 [12], H. chloromethanicum CM2 [18] and A. ciceronei IMB-1 (Harper et al., unpublished results)) exhibits halomethane:bisulfide/halide ion methyltransferase activity under aerobic conditions [7] has yet to be resolved. It is possible that both the methanethiol route (mediated by CmuA and possibly PaaE) and the H₄F route (mediated by CmuA and CmuB) operate under different physiological conditions in all of the organisms considered here. A role for the putative corrinoid methyltransferase, CmuC has yet to be defined. It could be involved in a third route (in addition to the H₄F and HS⁻ pathways) for transfer of methyl groups from CmuA to some, as yet unidentified acceptor, a role which would be consistent with speculation by Studer et al. [12]. The function of PaaE is also not clear but, in view of the homology with a family of oxidoreductases, it may exhibit the methanethiol oxidase activity observed in A. lissarensis CC495 [7].

Acknowledgements

This work was supported by funding from the UK Natural Environment Research Council for a studentship for KLW, and an Advanced Research Fellowship for IRM.