Structural homology modelling Intensive Phyre modelling was
Structural homology modelling. Intensive Phyre2 modelling  was performed using the primary amino bace inhibitor sequence of A1S_0222 as input to generate an atomistic 3D-homology model of A1S_0222. The fit to the SAXS data of the homology model as well as the fit to the data of the E. coli adenine-N6-DNA-methyltransferase, TaqI (Protein databank, PDB, accession 2ADM, chain A) were assessed using CRYSOL with 25 harmonics and a constant employed . Additional normal mode structural refinement of the A1S_0222 homology model was performed using SREFLEX .
Phylogenetic profiling. We determined the presence absence pattern of orthologs to A1S_0222 across 1548 A. baumannii strains, and 437 other taxa in the genus Acinetobacter. The ortholog search was performed in two stages. In stage one, we used the OMA algorithm  to compile an initial set of orthologous sequences from an all-against-all ortholog search in 104 representative Acinetobacter spp. representing all type strains. We then used this core orthologous group for a targeted ortholog search in the remaining taxa with HaMStR .
Results and discussion Motility assays. Preparing a library of A. baumannii ATCC 17978 mutants deficient in surface-associated motility, we found a knock out mutant of a putative DNA-(adenine N6)-methyltransferase, designated A1S_0222 in strain ATCC 17978. Fig. 1 represents the motility deficiency of the A1S_0222 mutant in strain ATCC 17978. This A1S_0222-specfic motility deficiency could be confirmed in other A. baumannii strains, for instance DSM 30011 (data not shown) and 29D2 (data not shown ). A1S_0222 protein overexpression and purification. To recombinantly produce the A1S_0222 protein, the a1s_0222 gene was amplified by PCR and cloned into the expression vector pGEX-6P-3 designed for expression of glutathione S-transferase (GST)-fusion proteins. The plasmid insert was confirmed by DNA sequencing and transformed into E. coli BL21 (DE3) pLysS for expression. Preliminary experiments revealed that the best conditions for protein expression are the induction with 0.05 mM IPTG and the incubation at 20 °C for 16 h (data not shown). The GST-A1S_0222 fusion protein was purified from a 200 mL E. coli culture and the soluble lysate was loaded on a GSTPrep™ FF 16/10 column. Eluted fractions were pooled and GST tag was cleaved off by PreScission protease. The protein was further purified using HiTrap™ SP XL column (cation exchange chromatography, CEC) and subsequently high-salt buffer (430 mM NaCl) was changed into a low-salt buffer (150 mM NaCl) using a PD-10 column. The molecular weight of recombinant A1S_0222 is about 49 kDa. The protein then was concentrated to about 2.5 mg/mL with the best concentration results being achieved with a Hydrosart membrane (10,000 Da MWCO) although concentration of the protein remained inefficient. The purification of A1S_0222 is illustrated in Fig. 2 and summarized in Table 1. We were able to purify A1S_0222 to near homogeneity. Identification of the A1S_0222 DNA recognition sequence and methyltransferase activity. According to REBASE, a database in which information about restriction enzymes, DNA methyltransferases, recognition and cleavage sites, published and unpublished references are collected [50,51], the A. baumannii methyltransferase A1S_0222 was found to be annotated as M.AbaBGORF222P and M.Aba17978ORF8565P recognizing the RAATTY DNA sequence. DNA adenine methyltransferases, sharing the same recognition site as A1S_0222 were found in various bacteria including the MTase M.Hpy300III (GAATTC) from Helicobacter pylori isolate BCM-300 and M.Mpa1757II (GAATTC) from Microcystis panniformis FACHB-1757 (http://rebase.neb.com/rebase/rebase.html). Unfortunately most m6A methyltransferases in this database are not yet biochemically characterized. To detect the recognition site of A1S_0222, Sanger DNA sequencing was performed (Fig. S1). In this automated dye terminator sequencing, m6A methylations in template DNA are identified by an increased complementary T signal . As a consequence the recognition site of A1S_0222 was detected to be GAATTC in which the second adenine (underlined) gets methylated. This recognition sequence equates to the restriction site of the endonuclease EcoRI (G↓AATTC). As the next step we confirmed the enzymatic activity of the purified m6A methyltransferase A1S_0222. Therefore a methylation/restriction protection assay was established (Fig. 3). A template-DNA (Int1), which contains only one A1S_0222 recognition site/EcoRI restriction site, was amplified by PCR. The purified PCR product Int1 was treated with A1S_0222 and afterwards incubated with and without EcoRI. As a control, Int1 was not treated with A1S_0222 but also incubated in the presence and absence of EcoRI. After incubation of Int1 with A1S_0222, EcoRI is no longer able to restrict the DNA Int1 (Fig. 3, lane “+/+”), whereas Int1 gets restricted into 2 fragments without the A1S_0222 treatment (Fig. 3, lane “−/+”). Thus, the results demonstrate the biological activity of the purified A1S_0222 as a DNA-(adenine N6)-methyltransferase. Note, that we did not observe any DNA degradation in this assay, demonstrating a DNase-free purification of A1S_0222. To address the specific activity of A1S_0222, a kinetic analysis of the methylation reaction was performed as described before. To this end, the methylation reaction was stopped and analyzed by agarose gel electrophoresis after 10 min, 20 min, 30 min and 40 min (Fig. 4). When incubating A1S_0222 for 10 min with 1 μg of DNA (Int1) half of the DNA was methylated and therefore protected from EcoRI restriction (lane “+/+”, Fig. 4). After 40 min 1 μg of DNA was completely methylated by A1S_0222 and could not be digested by EcoRI anymore. The specific activity of A1S_0222 was calculated to be approximately 0.013 mU/mg (Table 1). This low specific activity was confirmed using different DNA fragments (amplified from the A. baumannii 29D2 chromosome). One representative example is shown in Fig. S2. In this case, about 80% of DNA is methylated by A1S_0222 after 40 min of incubation. The specific activity of 0.013 mU/mg for A1S_0222 is low compared to other orphan methyltransferases. For example, the most prominent E. coli DNA adenine methyltransferase, Dam, shows a specific activity of about 8.9 × 105 U/mg . As general proof of the suitability of the assay conditions, all tested DNA fragments were also treated with a commercial EcoRI methyltransferase from E. coli RY13 (New England BioLabs). Fig. 4(B) and Figure S2(B) illustrate that EcoRI methyltransferase confers full protection from EcoRI restriction within 30 s of incubation.