br Materials and methods br Results br Discussion
Materials and methods
Discussion ParI, the first C5-DNA-MTase from a psychrophilic bacterium to be characterized, displays DNA methylation activity in vivo and in vitro. Its biophysical characteristics are consistent with large regions of disorder including thermal unfolding transitions over broad temperature range and low percentage of secondary structure as measured by CD. ParI unfolds at temperatures typical of many mesophilic proteins, but its temperature optimum could not be determined due to apparent loss of activity after purification. The in vivo assays demonstrate toxicity of ParI when recombinantly expressed in mcr+ E. coli at 37 °C, while assays in the native host were carried out at 25 °C, indicating that the enzyme is functional to some extent at both temperatures. Methyltransferase activity was also detected in vitro using radiolabelled 3H-SAM and pUC19 as a substrate. Activity was detected both in cell lysates and purified samples, but not after tag removal. The higher activity of His-MBP-ParI compared to His-ParI, coupled with the complete loss of activity upon MBP removal suggests that the higher specific activity is likely due to increased ParI solubility for this construct. An intriguing point is that, although untagged ParI lost activity after MBP removal and subsequent purification steps, it appears to retain its overall structure. ParI has PCQ-to-DCK substitution in motif IV compared with homologous MTases, where the proline has a role in orientation of the activated certainly and SAM in order to achieve methyl group transfer in the catalytic reaction . Alterations in this region may make ParI particularly prone to inactivation, while remaining in a somewhat folded state. Homology searches using BLAST revealed that ParI shares the highest sequence similarity (64%) to an uncharacterized C5-DNA-MTase from A. baumannii, which is also of phage origin and shares the motif IV PCQ-to-DCK substitution. The prevailing scenario for acquisition of orphan MTases is that during evolution, an MTase may have been part of a functional type II RM system that over time lost its REase member due to redundancy . MTases are generally well-conserved during evolution as they methylate at the same or an overlapping site as members of other type II RM system found in bacteria, thereby complementing methylation and protection of host DNA by these enzymes . A similar situation is found in the RM system of E. coli where Dcm, an orphan C5-DNA-MTase, has the same recognition sequence as the EcoRII RM system (CCWGG), and it is postulated that Dcm serves as back up for methylation of E. coli DNA [50,51]. Our hypothesis however, is that ParI was horizontally acquired as an orphan. Based on our bioinformatics investigations, as well as previously published data  we suggest that the gene encoding ParI is of phage origin, most likely a horizontally acquired gene from another bacterium during phage integration, with A. baumannii being a possible donor, although the original source still remains elusive. The parI gene is situated between two putative phage genes within the P. arcticus genome. The gene with the locus tag Psyc_0979, which is upstream to parI, is recognized by a BLAST homology search as part of temperate bacteriophage Psymv2 isolated from Psychrobacter sp. MV2, whose closest relative is an uncharacterized putative prophage within the P. arcticus 273–4 and A. baumannii genomes . The region downstream to parI, encoding the gene with the locus tag Psyc_0981 is similar to the MU phage . In addition to explaining the orphan status of ParI, horizontal gene transfer of parI into P. arcticus by a phage may explain why ParI does not exhibit typical cold-adapted features previously documented for other Psychrobacter enzymes such as the branched-chain 2-keto acid decarboxylase and murein peptide ligase which have optimal temperature of activity at 30 °C [52,53]. We also note that P. arcticus 273–4 is predicted to possesses an incomplete type IV McrBC-like REase, where the C subunit of the canonical BC complex were missing, making the McrB non-functional [54,55]. An evolutionary explanation for this may be that the substrate for a functional McrBC enzyme is methylated cytosine and, assuming that the ParI enzyme is functional, an active McrBC would restrict the methylated genomic DNA, leading to cell death. Therefore the McrBC-negative background of P. arcticus made it possible to retain a horizontally transferred methyltransferase. This would be consistent with the results of our activity assays, which showed both that ParI-methylated DNA is sensitive to cleavage by McrBC enzymes, and that transformation with a plasmid encoding ParI was lethal to a McrBC-containing bacterial strain. In addition, the C5-DNA-MTase deficient P. sp. PRwf-1 has a putative complete McrBC restriction system encoded in its genome.