Beyond a ribosomal RNA methyltransferase, the wider role of MraW in DNA methylation, motility and colonization in Escherichia coli O157:H7

MraW (RsmH) is an AdoMet-dependent 16S rRNA methyltransferase conserved in bacteria and plays a role in the fine-tuning of the ribosomal decoding center. It was recently found to contribute to the virulence of Staphylococcus aureus in host animals. In this study, we examined the function of MraW in Escherichia coli O157:H7 and found that deletion of mraW led to decreased motility and flagellar production. Whole-genome bisulfite sequencing showed genome wide decrease of methylation of 336 genes and 219 promoters in the mraW mutant. The methylation level of 4 flagellar gene sequences were further confirmed by bisulfite PCR sequencing. Quantitative reverse transcription PCR results indicated the transcription of these genes was also affected. MraW was observed to directly bind to the four flagellar gene sequences by electrophoretic mobility shift assay (EMSA). A common motif in differentially methylated regions of promoters and coding regions of the 4 flagellar genes was identified. Reduced methylation was correlated with altered expression of 21 of the 24 genes tested. DNA methylation activity of MraW was confirmed by DNA methyltransferase (DNMT) activity assay in vitro. The mraW mutant colonized poorer than wild type in mice. we further found that the expression of mraZ in the mraW mutant was increased confirming the antagonistic effect of mraW on mraZ. In conclusion, mraW was found to be a DNA methylase and has a wide-ranging effect on E. coli O157:H7 including motility and virulence in vivo via genome wide methylation and mraZ antagonism. IMPORTANCE MraW is a well-studied 16S rRNA methyltransferase and was recently found have an impact on bacterial virulence. Here we demonstrated its new function as a DNA methylase and effect on motility, colonization in mice, DNA methylation in genome wide and contribution to virulence. Its direct binding of differentially methylated flagellar-encoding DNA sequences was observed, indicating a correlation between DNA methylation and regulation of flagellar genes. In addition, the expression of mraZ which function as an antagonist of mraW was increased in the mraW mutant. mraW plays an important role in gene regulation likely through DNA methylation. Clearly it plays a role in virulence in E. coli O157:H7. It also opens a new research field for virulence study in bacteria.


104
DNA methylation has been well studied in eukaryote and is essential in the development 105 and progression of cancer (14). It becomes a rapidly growing area of research due to its 106 contribution to improved diagnosis and treatment. However, methylation effect on 107 virulence in bacteria has not been well studied and remains unknown. We hypothesized 108 that methylation in bacteria affect the virulence as it does in eukaryotes. The aim of this 109 study was to determine the effect of mraW on virulence in enterohaemorrhagic E. coli 110 6 O157:H7 and the relationship between methylation and virulence.

112
Deletion of mraW led to reduced motility and flagellin production/secretion. The 113 gene encoding MraW was deleted in E. coli O157:H7 strain EDL933 resulting in a 114 mutant strain designated as EDL933ΔmraW. The motility of the wild type and the 115 mutant was determined by the radius of chemotactic ring which was 2.3 ± 0.1 cm and 116 0.783 ± 0.03cm respectively. The difference is statistically significant (t test, P < 0.01) 117 ( Fig.1.A-B). The decrease in motility in the mraW mutant was also correlated with a 118 decrease in the production or secretion of FliC as determined by western blotting using 119 anti-H7 antisera (Fig. 1C). To determine whether the mraW deletion can be 120 complemented, we created pBADmraW, a low copy number plasmid carrying mraW 121 which was transformed into EDL933ΔmraW. The radius of the chemotactic ring of the 122 complemented strain (EDL933ΔmraW+pBADmraW) was 2.55 ± 0.09 cm (Fig. 1A), 123 which is similar to the wild type. Thus, the decrease in motility of the mutant was almost 124 complemented back by the plasmid pBADmraW expressing mraW. The generation 125 time of EDL933, EDL933ΔmraW and EDL933ΔmraW+pBADmraW was 35.7 ± 0.06 126 min, 35.9 ± 0.04 min and 35.8 ± 0.1 min respectively. Hence the difference in motility 127 and production of FliC was not due to the growth rate which was similar among the 128 mutant, the complemented strain and the wild type. 129 To investigate whether the reduced motility of the EDL933mraW mutant was due to   Table S1 in the 149 supplemental material ). Although a similar methylation level was found in cytosine in 150 EDL933ΔmraW compared to EDL933, there was a trend of differences in methylation  (Table S2 and S3).  Many virulence genes and promoters in the mraW mutant were also affected at 177 methylation level (Table S2 and (Table S4). containing 6 DMRs while fliR had no DMR. One motif, GATGAAAGGC, was 202 common in these 6 DMRs with a p value < 0.0001 (Fig. 3A). Similarly, common motifs 203 in 21 DMRs of the three heat shock protein genes htrC, ddg and grpE were investigated.

205
ATTACT occurs 1330 times in the EDL933 genome, we did not pursue this motif  (Table S5). Three protein concentrations were tested. Binding occurred at the lowest  were found at day 1. A significant difference in colonization between EDL933 and 245 EDL933ΔmraW was found from day 2 to day 7 ( Fig. 4B-C). A 100 times difference 246 was found between the wild type and the mraW deletion mutant from day 2 in both 10 9 247 and 10 10 CFU inoculum groups. The results indicated that the mraW deletion mutant 248 colonized the mice poorer than the wild type EDL933. encoding heat shock proteins, htrC, ddg and grpE, were significantly increased at the 258 transcriptional level (t test, P < 0.01) (Fig. 6B). The mRNA transcription of cold shock 259 protein cspA was also increased in a low but significant level (t test, P < 0.01) (Fig. 5C).

263
Type III secretion phenotype was investigated in the mutant. T3SS proteins were 264 13 compared between the wild type and the mutant. No significant difference was found 265 in protein level between these two strains (Fig.5D). Although the transcription levels of 266 escD, espB and z4187 were all increased (Fig. 5D).

267
Due to the antagonistic functions of mraW and mraZ and their effect on virulence, the 268 expression of mraZ was also examined. mraZ expression showed more than 7 times 269 increase in the mraW mutant compared to the wild type, confirming the antagonistic 270 effect previously observed (19) (Fig. 5D).   Although the vast majority of the differentially methylated genes/promoters had 359 reduced methylation in the mraW mutnat, 13 showed increased methylation levels. A 360 plausible explanation for these small number of genes with increased methylation is 361 that these DMRs may be methylated by other methylases that were repressed by mraW. 362 It is unlikely all of these were false positives as two were confirmed by BSP PCR 363 sequencing. Further studies will be required to elucidate the mechanisms involved.

364
In conclusion, we found that mraW plays a role in gene regulation through DNA  Table S4. Bacteria were routinely cultured in    (Table S4). Primers including 5-388 mraWF and 3-mraWR were used (Table S5). E. coli strain DH5a was used as the 389 intermediate host strain for cloning and all constructs were verified by sequencing.

390
Motility assays. Swimming motility was evaluated as described by Xu and Xu (7).

391
Briefly, wild type, the mraW mutant and complement strains were cultured overnight 392 and stab inoculated with a sterile inoculating needle and incubated at 37°C for 16h. All 393 strains were tested in triplicate and each experiment was carried out on three separate 394 occasions. The motility radius of each strain was measured and analyzed by t-Test.

395
Transmission electron microscopy. TEM was carried out as previously described (7).  (Table S5, from fhiAP1F to treR2R). PCR was carried out using the DNA template 435 which was treated by EpiTect ® Bisulfite kit. PCR product was cloned into pMD-18-T 436 and sequenced. For each DNA fragment, at least 10 clones were selected for sequencing 437 to minimize the sequencing error. Cytosine methylation data was achieved by mapping 438 the sequencing data to the reference E. coli O157:H7 str. EDL933.

439
RNA extraction, cDNA synthesis and quantitative reverse transcription PCR.