E. coli MV1161, a derivative of the standard laboratory strain AB1157, was used as reference strain in this work. Isogenic ΔnanA::kan mutant was constructed using the λ Red technique (Datsenko and Wanner, 2000). The MV1161 ΔnanA::kan mutant was transformed with pMAL-c5x-nanA plasmid containing the nanA gene in order to complement lyase activity. Both the ΔnanA mutant and the complement strain showed a growth rate superimposible with the parental strain. The strain LF82 (Boudeau et al., 1999), a prototype adherent-invasive E. coli (AIEC) strain, was used in adhesion and invasion experiments with Caco-2 eukaryotic cell line (http://www.lgcstandards-atcc.org/products/all/HTB-37.aspx?geo_country=it). LF82, was grown overnight in LB medium at 37°C, 180 rpm agitation, then used to inoculate 1:100 two different 250 mL flasks containing 100 mL of 22°C pre-warmed LB and underwent growth at 37°C, 180 rpm. Once reached an OD₆₀₀ = 0.5, MMS was added to one flask at 0.04%, and OD₆₀₀ measurements were taken every h from both flasks.

DIGE experiments were performed on four biological replicates of MMS treated and four biological replicates of MMS untreated E. coli as previously described (Caterino et al., 2009). E. coli cells were homogenized in 0.5 ml of lysis buffer (7 M urea, 2 M thiourea, 4% chaps, 30 mM Tris-HCl pH 7.5) using a Dounce homogenizer. Concentration of the protein extracts was determined and equal amounts of protein lysates were then labeled in vitro using two different fluorescent cyanine minimal dyes (Cy3 and Cy5, respectively) differing in their excitation and emission wavelengths. A third cyanine dye (Cy2) was used to label a mixture of all samples as internal standard. The three differently labeled protein mixtures were pooled and subjected to isoelectric focusing on a 3–10 pH range with 24 cm strips using an IPGphor 2 (GE Healthcare) at 20°C (costant voltage 300 V for 3 h, voltage gradient to 1000 V in 6 h, voltage gradient to 8000 V in 3 h and costant voltage 8000 V for up to 4 h). The current limit was 50 mA/strip. The samples for the semipreparative gels were focused in a separate run using the same IEF condition. Following IEF, strips were equilibrated in 100 mM Tris, pH 8.0, 30% v/v glycerol, 2% w/v SDS, 6 M urea, reduced with 0.5% w/v DTT for 15 min and alkylated with 4.5% w/v iodoacetamide for further 15 min. IPG strips were then transferred to the top of a classical 12.5% SDS PAGE gel for a second orthogonal electrophoresis analysis. The SDS PAGE separation was performed at 2 W/gel, for 16 h at 25°C.

The Cy2, Cy3, and Cy5 images were obtained by scanning each of the four DIGE gels at excitation/emission wavelength of 480/530 nm for Cy2, 520/590 nm for Cy3 and 620/680 nm for Cy5 using a Typhoon 9410 TM scanner (GE Healthcare). The semi-preparative gel, prepared in an identical fashion, was scanned with 480/633 nm wavelengths. After consecutive excitation at both wavelengths, the images from the preparative gel were overlaid and subtracted (normalized) from the samples, whereby only differences (up or down expressed proteins) between the two samples were visualized. DeCyder software 5.01 (GE Healthcare) was used to compare the abundance changes between the control and treated samples. The Batch Processor Module (GE Healthcare) was used for normalization of the signals from each CyDye channel and spot detection on the 12 images from the four DIGE gels and the calculation of the volume ratio for each spot pair (Cy3:Cy2 and Cy5:Cy2 ratios). The Biological Variation.

Analysis (BVA) module (GE Healthcare) was used to match all the 12 spot maps from the four DIGE gels and calculate the average abundance changes and paired Student's t-test p-values for the variance of these ratios for each protein pair across all four samples.

By performing a high resolution image analysis on the four biological replicates, it was possible to visualize significant differences between numerous protein spots present on the gels. Differential spots were defined as having a variation higher than 1.2 (p < 0.05) per previously established methods (Cella et al., 2012). The gels showed a high degree of similarity, with more than 80% of all spots superimposeable. The remaining 20% showed variation and were further studied.

The spots of interest were excised, hydrolyzed and the peptide mixtures analyzed by MALDI-MS and LC-MSMS mass spectrometry, on a 4800 Plus MALDI TOF/TOF™ (Applied Biosystems, Framingham, MA, USA) and a LC/MSD Trap XCT Ultra (Agilent Technologies, Palo Alto, CA) equipped with a 1100 HPLC system and a chip cube (Agilent Technologies) respectively. MALDI spectra were acquired in the positive ion reflector mode using delayed extraction in the 800–4000 Da mass range. LC-MS/MS analysis was performed using data-dependent acquisition of one MS scan followed by MS/MS scans of the three most abundant ions in each MS scan. Raw data analyses were converted into a Mascot format text to identify proteins using Matrix Science software. The protein search considered the following parameters: non-redundant protein sequence database (NCBInr), specificity of the proteolytic enzyme used for the hydrolysis (trypsin), taxonomic category of the sample, up to one missed cleavage, cysteines as S-carbamidomethylcysteines, unmodified N- and C-terminal ends, methionines both unmodified and oxidized, putative pyro-Glu formation by Gln, precursor peptide maximum mass tolerance of 200 ppm, and a maximum fragment mass tolerance of 200 ppm.

Identified proteins were investigated to predict functional protein association networks for each entry using the STRING online database (http://string-db.org). Proteins were analyzed with the STRING software.

Three E. coli culture replicates were grown at 37°C in the presence and in the absence of 0.04% MMS according to proteomics experiments. In order to obtain an immediate stabilization of RNA, two volumes of RNA protect Bacteria Reagent (Qiagen) were added directly to one volume of bacterial culture (10 OD600 nm). Bacterial cells were harvested and each pellet was stored at -20°C overnight. Total RNA was extracted from culture using RNasy Midi Kit (Qiagen) according to the manufacturer's instructions. The purity and integrity of total RNA was verified by both 1% denatured agarose electrophoresis and by evaluating that the A260/A280 absorbance ratio was within the range 2.1–2.3.

Total RNA (1 μg) was reverse transcribed in 40 μL final reaction volume using TaqMan Universal PCR Master Mix kit (Applied Biosystem) according to the manufacturer's instructions. The cDNA was synthesized at 48°C for 60 min and then stored at −20°C. Real-time PCR primer sequences were designed using Primer Express 3.0 software. Quantitative real-time polymerase chain reaction (qRT-PCR) was performed using SYBR green PCR Master Mix and Step ONE Real Time PCR System (Applied Biosystems) according to the manufacturer's instructions to measure nanA gene expression. rpoA amplifications were used as normalization controls. Each sample was analyzed three times to obtain average data. Relative transcript levels were calculated using the 2^−ΔΔCt formula (Ct = threshold cycle; Livak and Schmittgen, 2001).

E. coli MV1161 was amplified from host DNA by polymerase chain reaction (PCR) using the following primers: FW AGCGGATCCATGGCAACGAATTT; RW AATAAGCTTTCACCCGCGCTCTT, where the recognition sites for BamHI and HindIII are underlined. To obtain the nanA gene, the PCR product was digested with the appropriate restriction enzyme and cloned into the BamHI and HindIII sites of the pET28a vector. The resulting plasmid contained the coding sequence for the recombinant NanA protein fused to a 6-histidine tag to facilitate protein purification.

The recombinant gene was expressed into the E. coli strain BL21. Cells were grown in LB medium at 37°C with 50 μg/mL kanamicin to an optical density of ~0.5 at 600 nm, and 0.1 mM isopropyl-beta-D-thiogalactopyranoside (IPTG) was added. The culture was grown for 3 h at 37°C for NanA production. Cells were harvested by centrifugation at 5000 rpm for 15 min at 4°C. The cellular pellet was resuspended in equilibration buffer (20 mM Na₂HPO₄, pH 7.4, 500 mM NaCl, 20 mM imidazole) containing 1 mM PMSF and disrupted by passage through a French Press.

The cell extract was centrifuged at 13000 rpm for 20 min at 4°C, to remove cell debris and the supernatant was filtered with a syringe-driven filter (0.22 μm) before protein purification. Soluble cell extract was loaded onto a His-Select Nickel Affinity Gel equilibrated with equilibration buffer and the bound protein was eluted with 500 mM imidazole in 50 mM phosphate buffer pH 7.4, 500 mM NaCl.

Protein concentration was estimated with Bradford reagent (Bio-Rad protein assay), protein purity was assessed by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and its primary structure verified by MALDI mapping strategy. Purified NanA protein was used to produce antiNanA antibodies by rabbit immunization (PRIMM, Milano, Italy).

E. coli cultures were grown at 37°C in the presence and in the absence of 0.04% MMS according to proteomics experiments. The cell pellets were lysed by French Press, resolved on a 12.5% SDS-PAGE gel and transferred onto a PVDF membrane. The membrane was blocked with 3% v/v milk in PBS buffer at room temperature for 1 h. The immunodetection was conducted by incubating the membrane with the anti-NanA antibodies (1:300, Primm, Milano Italy) for 18 h at 4° C under slow stirring. The membrane was then incubated with the secondary antibody conjugated to horseradish peroxidase (1:50000) for 1 h. Target proteins were visualized by ECL detection (Pierce).

A quantity of 2 mL of overnight bacterial cultures grown in LB + 1% glucose was added into 6-well flat bottomed polystyrene plate (Falcon). The plates were incubated at 37°C in aerobic conditions in the presence of 0.04% MMS. Growth was monitored by measuring the OD600, and after 24 h incubation the ability of the E. coli strain to adhere to the polystyrene plates was tested. The liquid medium was removed and the wells washed with sterile distilled water. The plates were then stained with 0.1% crystal violet for 20 min (O'Toole, 2011). Excess stain was rinsed off by placing the plate under running tap water. After the plates were air dried, the dye bound to the adherent cells was solubilized with 20% (v/v) glacial acetic acid and 80% (v/v) ethanol per well. The OD of each well was measured at 590 nm.

Biofilm assays on the complemented null NanA mutant strain were performed in the same conditions described above in the presence of 0.1 mM IPTG. Biofilm formation in the presence of a synthetic inhibitor was carried out on E. coli MV1161 strain grown in sterile 6-well flat-bottomed polystyrene plates treated with different concentrations of N-Acetyl-2,3-dehydro-2-deoxyneuraminic acid (DANA) ranging from 1 to 100 μg/mL. The plates were incubated aerobically at 37°C for 24 h and biofilm production was quantified as previously described.

Bacteria from 18 h cultures in BHI broth, grown in the absence of MMS were further subcultured up to OD600 of 0.5 at 37°C with or without MMS 0.04%. HeLa cells, cultured in 24-well plates to obtain semi-confluent monolayers (1 × 10⁵ cells/well), were infected with 0.05 mL of logarithmically grown bacteria for 1 h at 37°C.

For the invasion assay, the monolayers were washed with PBS and treated with 0.5 mL of fresh medium containing 200 mg/mL of gentamicin for 1 h at 37°C to kill extracellular bacteria. Cells were then lysed by the addition of 0.025% Triton X-100 and plated on LB to count viable intracellular bacteria (Artini et al., 2012).

For the adhesion assay, loosely bound bacteria were removed from the cell monolayers by two washes with PBS. Cells were then lysed with 0.025% Triton X-100 and plated on LB agar to determine viable adherent bacteria. Adhesion efficiency was expressed as the percentage of the inoculated bacteria that adhered to HeLa cells.

Caco-2 intestinal cells were used for invasion and adhesion assays with AIEC LF82 strain (Darfeuille-Michaud et al., 2004). Caco-2 cells were seeded in a 12-well plate at a density of 4 × 10⁵ cells/well, let stand for 1 day, and medium (DMEM with low glucose supplemented with 10% FBS) replaced the day of infection. One mL of a 6 h-old LF82 culture (logarithmic phase) was harvested, and MMS at 0.04% was added for 20 min in one tube, while another was left non-treated as a control. After the induction step, both tubes were centrifuged at 3000 rpm for 5 min, and bacterial pellets resuspended in PBS. 10 μL of resuspended pellets were added to each well to obtain a MOI of bacteria/cells 10:1, and 12-well plates were incubated for 4 h at 37°C, 5% of CO₂.

For the invasion assay, cells were washed three times with PBS and 2 mL of 22°C pre-warmed culture medium supplemented with 400 μg/mL of gentamicin were added to each well and the plate was incubated for 2 h at 37°C, 5% of CO₂. Gentamicin was then removed washing three times with PBS and 1 mL of 1% Triton X-100 was added for 5 min during which the plates were gently swirled three times. Serial dilutions in PBS were performed till 10⁻⁶ and 100 μL of each dilution were plated onto TSA petri dishes (repeated three times). Colony-forming units (CFU) per mL were enumerated and the invasion index was expressed as the number of invasive bacteria per single Caco-2 cell.

For the adhesion assay, cells were washed three times with PBS and 1 mL of 1% Triton X-100 was added for 5 min, during which the plates were gently swirled three times. This concentration of Triton X-100 had no effect on bacterial viability for at least 30 min. Serial dilutions in PBS were performed till 10⁻⁶ and 100 μL of each dilution were plated onto TSA petri dishes (repeated three times). Colony-forming units (CFU) per mL were enumerated and the adhesion index was expressed as the number of adhesive bacteria per single Caco-2 cell.

Three biological replicates of both the invasion and adhesion assays were performed and the results expressed as mean ± standard error of the mean.

Differential in Gel Electrophoresis (DIGE) was used to examine the global effect of methylation stress on E. coli. Bacterial cells were grown in four replicates and each culture was divided into two aliquots, one of them was treated with 0.04% MMS and the other one was kept untreated and used as control. Protein lysates from MMS treated and untreated cells were extracted and fractionated by bidimensional electrophoresis using a 3–10 pH gradient. The relative quantitative analysis was performed so that only proteins up or down expressed in all four treated cell samples versus control were considered (Figure 1A). The corresponding preparative gel is reported in Figure 1B, where differently expressed proteins in the presence of MMS are labeled and were submitted to the identification procedure.

The spots of interest were picked, in situ hydrolyzed with trypsin and the resulting peptide mixtures analyzed by LC-MS/MS. Proteins were identified using the Mascot software. A total of 69 differentially expressed proteins were identified in MMS treated E. coli cells compared to controls, 61 down and 8 up expressed, and are reported in Table 1. Details of mass spectrometry analysis were reported in Table 1S of Supplemental Materials.

The network distributions of the 69 differently expressed proteins were explored using STRING software. The top-ranked networks were in Metabolic Pathways with a p-value of 3.45E-10, Glycolysis / Gluconeogenesis (p = 5.87E-8) and Microbial metabolism in diverse environments (p = 2.11E-7). Stronger associations are represented by thicker lines (Figure 1C).

Most of the identified proteins were under expressed with only a few found over expressed. The majority of the down expressed proteins belong to metabolic processes, suggesting a decrease in energy production. However, the most intriguing result in the presence of MMS was the sharp decrease in the expression of the N-acetylneuraminate lyase (NanA) an enzyme involved in sialic acid metabolism. Under methylating stress conditions, NanA expression was reduced by a factor of 2⁶, which corresponds to about 60 times decrease in protein concentration.

Proteomic data were independently confirmed by western blot analysis. Recombinant NanA protein bearing a N-terminal His tag was homologously produced in E. coli and purified by metal affinity chromatography. The recombinant protein was characterized by SDS-PAGE and its primary structure verified by MALDI mapping (Figure 1S of Supplemental Material). Specific anti NanA antibodies were then produced in rabbit, tested using different amounts of E. coli extracts and used to assess the expression of NanA under methylation stress conditions.

Figure 2A shows the western blot analysis of protein extracts from E. coli grown in the absence and in the presence of 0.04% MMS; the corresponding densitometric analysis is shown in Figure 2B. Protein content was normalized using the Maltose Binding Protein that was demonstrated to be unaffected by methylation stress by DIGE analysis. A strong decrease in the amount of the lyase was clearly detected upon methylation stress, thus confirming proteomic results.

The significant down regulation of NanA was further investigated at mRNA level. Total RNA was extracted from E. coli cells either treated or untreated with 0.04% MMS demonstrating a high quality product.

Total RNA was reverse transcribed to obtain cDNA. Real time PCR assays were carried out to detect nanA gene expression levels in both MMS treated and untreated E. coli cells. Melting point curves displayed a single peak and any amplification of non-specific targets or primer dimers was not observed demonstrating the specificity of amplification. rpoA was used as housekeeping gene to normalize the threshold cycle values (Ct). Analyses were performed in triplicate.

The results of PCR amplification are shown in Figure 2C. Transcription of nanA gene was unaffected by MMS treatment showing a RNA amount almost similar to the untreated cells being the difference shown in Figure 2C not significant. Altogether these data suggested the occurrence of a negative translational control of nanA exerted by an unknown cellular mechanism in the presence of methylation stress.

Next, we explored the possibility that decreasing NanA levels upon exposure to methylating agents might also affect biofilm formation and adhesion abilities in the non-pathogenic MV1161 E. coli strain.

E. coli cells were then grown in LB medium supplemented with 1% glucose either in the presence or in the absence of 0.04% MMS. As shown in Figure 3A, methylation stress decreased biofilm formation by about 60% compared to non-treated cells.

Since biofilm mechanisms are usually associated to cellular adhesion/invasion processes, we tested the effect of methylation stress on the capabilities of MV1161 to adhere to human HeLa cells. Following 0.04% MMS treatment, E. coli cells were subcultured and several different dilutions added to HeLa cells in order to establish the best bacterial/eukaryotic cells ratio. Analyses were performed in triplicate. Data obtained for one of the three cellular dilutions are reported in Figure 3B showing that E. coli adhesion on HeLa cells was drastically decreased in the presence of 0.04% MMS.

Since NanA expression was heavily decreased following MMS treatment, these data might suggest a possible involvement of this enzyme in the bacterial cohesive properties affecting the biological processes leading to biofilm formation.

These observations prompted us to further investigate the putative role of NanA in the adhesion properties of E. coli. A null NanA mutant was constructed transforming electrocompetent E. coli MV1161 strain (Datsenko and Wanner, 2000). Biofilm formation experiments were carried out on ΔnanA mutant in comparison with MMS treated and untreated MV1161 E. coli cells. As reported in Figure 4, biofilm formation was drastically reduced in the ΔnanA mutant with the effect being very similar to that observed for the MMS treated MV1161 strain.

Next we checked the effect of MMS on the residual biofilm formation in the nanA mutant strain to investigate whether the observed decrease in the MMS treated cells and in the nanA mutant strain are both due to loss of NanA. Biofilm formation experiments were performed on ΔnanA mutant under methylation stress conditions in comparison with untreated MV1161 strain as control. Treatment with MMS did not reduce further the biofilm levels in the nanA mutant strain as shown in Figure 4, thus indicating that the two effects were not additive.

The biological implication of the lyase in biofilm formation was further supported by complementation assays. The ΔnanA mutant was transformed with a plasmid vector harboring a wild type copy of the gene, and the resulting complemented strain was used in biofilm assays as previously described. Figure 4 shows that biofilm production was greatly increased in the complement strain in comparison to MV1161 wild type strain, demonstrating that re-integration of the lyase NanA fully restored the ability of the mutant strain to adhere to abiotic surface.

The fundamental role of NanA in biofilm formation was finally assessed by performing biofilm experiments in the presence of N-Acetyl-2,3-dehydro-2-deoxyneuraminic acid (DANA). Although DANA is known as a sialidase inhibitor, we hypothesized that, being structurally related to sialic acid, it might also act as a competitive inhibitor of the lyase activity. Biofilm formation was performed in the presence of increasing DANA concentrations ranging from 1 to 100 μg/mL. Figure 5 shows biofilm production of E. coli MV1161 strain in the absence and in the presence of 50 μg/mL DANA inhibitor. From these experiments, we can infer that inhibition of lyase activity greatly affected the biofilm production properties of E. coli as demonstrated by the clear decrease in biofilm formation to a level very similar to that observed following MMS incubation.

In order to verify if MMS could affect adhesive and invasive properties of a pathogenic E. coli, we used the LF82 strain, an adherent-invasive E. coli (AIEC) pathogen associated with Crohn's disease development (Boudeau et al., 1999). This strain showed strong adhesive and invasive capabilities in intestinal cells and prolonged intracellular survival in macrophages, leading to enhanced pro-inflammatory cytokines release (Bringer et al., 2012).

As shown in Figure 6A, upon addition of 0.04% MMS, LF82 strain showed a decreased steady state level in the growth curve (OD_600LF82 = 4.7 vs. OD_600LF82+MMS = 1.6), a different mid-point (t_{mid−pointLF82} = 5.9 h vs. t_{mid−pointLF82+MMS} = 3.8 h), and a diminished intrinsic growth rate in the logarithmic phase (r_LF82 = 0.76 vs. r_LF82+MMS = 0.37). This result revealed that MMS affected LF82 growth curve, leading to the hypothesis of a parallel alteration in its adhesive and invasive properties.

To asses this point, we performed adhesion and invasion assays upon 4 h of LF82 infection on intestinal Caco-2 cells, a cell line deriving from gut epithelium, the initial target of LF82 adhesion and invasion, in the presence of MMS at concentrations of 0.04%, which did not affect the viability of the pathogenic strain (Figure 6B). Results showed that adhesion was diminished by 60.2% (Mann-Whitney, p = 0.0238), while invasion by 53.5% (Mann-Whitney, p = 0.0353; Figure 6C). Interestingly, MMS treatment led also to a diminished aggregative behavior of LF82, which usually aggregates within cell-to-cell junctions on Caco-2 monolayers (data not shown).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.