Structural characterization of A S Given the low
Structural characterization of A1S_0222. Given the low specific activity of our protein preparations and the considerable problems to concentrate the protein to levels above 2.5 mg/mL, we employed SAXS to structurally characterize and generate a low-resolution model of A1S_0222 in solution. Fig. 5 and Table 2 summarize the SAXS results which demonstrate that the protein is monomeric. Both concentration-dependent MW (52 kDa) and concentration-independent MW estimates (45–60 kDa) are in the range of the expected MW calculated from the amino self-flagellation sequence for the monomer (49 kDa). The shape classification of the data and the resulting p(r) vs r profile indicate that A1S_0222 forms a compact, slightly anisotropic structure with an R of ∼3 nm, with a D of ∼11 nm (Table 2). The consensus low-resolution (2.5 nm) DAMMIN bead model (χ = 1.06; CorMap p = 0.13) shows that A1S_0222 is flattened in one dimension, with possible structural extensions from the predominantly compact core (refer to SASBDB entry SASDD32). Of interest, a simple search of the PDB for DNA methyltransferases and subsequent fitting of the atomic structure of the E.coli adenine-N6-DNA-methyltransferase, TaqI (PDB 2ADM chain A ) to the SAXS data suggests that TaqI and A1S_0222 may share structural features in common (Fig. 6). Although the TaqI structure does not fit the SAXS data (χ2 = 2.5, CorMap p = 0), the modelled intensities of the TaqI methyltransferase are not totally dissimilar across the experimental A1S_0222 SAXS profile which is surprising considering the two proteins only share 17% amino acid sequence identity. Indeed subsequent Phyre2 homology modelling of A1S_0222, using the amino acid sequence of the protein as a template, consistently identified a number (more than 20) of high-scoring homologous fragments in the PDB (with greater than 97% confidence) that are categorized as DNA methyltransferases. The top-scoring homologues identified during the Phyre2 modelling are that of E. coli DNA adenine methyltransferase, PDB 2G1P (chain B) and bacteriophage T4 DNA adenine methyltransferase, PDB 1YF3 (chain A). As with TaqI, the final A1S_0222 Phyre2 homology model shows significant statistical discrepancies when fitting the model to the SAXS data (χ2 = 2.6; CorMap p = 0). However, if the Phyre2 A1S_0222 homology model undergoes additional normal mode rigid-body refinement – that allows for subtle shifts in domain orientation – the predicted structure effectively ‘opens up’ and undergoes an overall extension to subsequently fit the SAXS data (χ2 = 1.2; CorMap p = 0.019, Fig. 6).
. Lastly, we analyzed the phylogenetic distribution of A1S_0222 within Acinetobacter spp.. Our analysis revealed that this protein is encoded in all but 19 of the 1985 genomes covering the full known diversity of the genus Acinetobacter. The 19 cases where no ortholog could be identified are shown in Table 3. We see no obvious grouping of individual species where this gene is missing. Instead, we conclude that this protein is ubiquitously present in Acinetobacter spp.. Most probably, issues connected to data quality and the draft status of most Acinetobacter spp. genome reconstructions explain the few cases where no ortholog could be detected.
Conclusion In conclusion, we confirmed the classification of A1S_0222 as a Type II N6-adenine DNA methyltransferase recognizing GAATTC in line with the REBASE in silico classification . A1S_0222 seems to act as an orphan methyltransferase since no evidence for an associated endonuclease could be found in REBASE. We propose the name AamA (adenine methyltransferase A) in addition to the formal names M.AbaBGORF222P and M.Aba17978ORF8565P provided by REBASE. To the best of our knowledge, this is the first DNA methyltransferase of the nosocomial pathogen A. baumannii that has been experimentally studied.