Mutations in hpyAVIBM, C5 cytosine DNA methyltransferase from Helicobacter pylori result in relaxed specificity

Authors


D. N. Rao, Department of Biochemistry, Indian Institute of Science, Bangalore 560 012, India
Fax: +91 80 2360 0814
Tel: +91 80 2293 2538
E-mail: dnrao@biochem.iisc.ernet.in

Abstract

The genome of Helicobacter pylori is rich in restriction–modification (RM) systems. Approximately 4% of the genome codes for components of RM systems. hpyAVIBM, which codes for a phase-variable C5 cytosine methyltransferase (MTase) from H. pylori, lacks a cognate restriction enzyme. Over-expression of M.HpyAVIB in Escherichia coli enhances the rate of mutations. However, when the catalytically inactive F9N or C82W mutants of M.HpyAVIB were expressed in E. coli, mutations were not observed. The M.HpyAVIB gene itself was mutated to give rise to different variants of the MTase. M.HpyAVIB variants were purified and differences in kinetic properties and specificity were observed. Intriguingly, purified MTase variants showed relaxed substrate specificity. Homologues of hpyAVIBM homologues amplified and sequenced from different clinical isolates showed similar variations in sequence. Thus, hpyAVIBM presents an interesting example of allelic variations in H. pylori where changes in the nucleotide sequence result in proteins with new properties.

Abbreviations
AdoMet

S-adenosyl-l-methionine

IPTG

isopropyl thio-β-d-galactoside

MTase

methyltransferase

Introduction

DNA methylation has been suggested to play multiple roles in prokaryotes and eukaryotes [1,2]. DNA methylation plays a critical role in bacteria with respect to protection from ‘foreign’ invaders such as bacteriophages and transposons [3]. In prokaryotes, methylation can occur at N6 adenine, C5 cytosine and N4 cytosine [1]. It was previously reported that N6 adenine methylation plays a significant role in gene regulation in a number of bacteria [4,5]. Two solitary adenine methyltransferases (MTase), Dam and CcrM, are known to be involved in epigenetic regulation in a number of bacterial species [2]. On the other hand, the role of cytosine methylation is not very well established. In some bacteria, cytosine-MTase (Dcm) is associated with very short patch repair (Vsr) endonucleases [6]. It is also known that the methylation of cytosine by DNA MTase plays a critical role in the high frequency of C→T or corresponding G→A transition mutations [7,8]. Thus, methylation of DNA cytosine residues is highly mutagenic. Methylated cytosines were first found to comprise base substitution hot spots in Escherichia coli. Various studies have suggested that transition mutations in bacteria are the result of elevated spontaneous deamination of methylated cytosine to thymine [9]. The hydrolytic deamination rate of C5 methyl cytosine is two- to three-fold higher than that of C in double-stranded DNA [10] and four- to five-fold higher in single-stranded DNA [11,12]. The presence of the HpaII MTase in E. coli was shown to cause a substantial increase in C→T mutations at CG sites [13]. E. coli MTase Dcm had a similar mutagenic effect within its own recognition sequence (5′-CCWGG-3′) [9].

Helicobacter pylori is an obligate parasite of humans and the genomic sequences of several H. pylori strains have revealed that this bacterium contains a number of restriction–modification (RM) genes, some of which have been subsequently shown to function as authentic RM systems or as partial systems with only the DNA MTase component being active [14,15]. Interestingly, RM genes comprise a significant percentage of H. pylori strain-specific genes [15]. The identification and study of both species-specific and strain-specific MTases of H. pylori could enhance our understanding of the pathogenic mechanisms of this organism.

M.HpyAVIB is a C5 cytosine MTase from H. pylori [14]. In H. pylori strain 26695, hpyAVIBM exists as an overlapping ORF with another MTase hpyAVIAM. These MTases are possibly remnant MTases of a defunct RM system. Both these ORFs have a high sequence identity with the MnlI DNA MTase, part of a Type IIS RM system [16]. However, in H. pylori, the functional MnlI restriction enzyme is absent [17]. The recognition sequence of M.HpyAVIB was previously reported by Vitkute et al. [18] based on restriction enzyme digestion as 5′-CCTC-3′, where C is methylated by M.HpyAVIB in the target site. hpyAVIBM has a stretch of dinucleotide repeats (AG), which makes it a candidate for phase variation [19]. Phase variation plays a vital role in a number of pathogenic bacteria where it is used to facilitate immune evasion in a host [20].

The present study highlights the effects of DNA methylation by M.HpyAVIB, especially the increased mutation rates, resulting in an accelerated micro-evolution of the MTase gene.

Results

Oligomeric status of M.HpyAVIB protein

M.HpyAVIB protein was purified to near homogeneity as described in the Experimental procedures (Fig. S1A). M.HpyAVIB protein eluted as a monomer with a molecular weight of 44 kDa as determined by analytical gel filtration chromatography (Fig. S1B). In addition, the molecular weight of the protein was determined by MALDI to be 43.47 kDa (Fig. 1A). The hydrodynamic radius (Rh) of M.HpyAVIB was calculated as 2.8 nm by dynamic light-scattering (Fig. 1B). The frictional ratio was calculated as 0.92, indicating that M.HpyAVIB is more or less spherical in structure (an ideal spherical protein would give a value of 1.0) [21].

Figure 1.

 (A) Molecular mass determination of M.HpyAVIA protein by MALDI. (B) Light scattering was performed with 20 μL of M.HpyAVIB protein (1.5 mg·mL−1) in 10 mm Tris–HCl (pH 8.0), 100 mm NaCl, 5 mmβ-mercaptoethanol using dynamic light-scattering. (C) Differential effect of metal ions on the methylation activity of wild-type M.HpyAVIB and (D) the M.HpyAVIB variant. (1) In the absence of any metal ion; (2) in the presence of 4 mm EGTA; (3) in the presence of 2 mm EDTA; (4) in the presence of 1 mm Ni2+; (5) in the presence of 1 mm Ni2+ and 2 mm EDTA; (6) in the presence of 2 mm Ca2+; (7) in the presence of 2 mm Ca2+ and 4 mM EGTA.

Kinetics of the methylation reaction

To establish the relationship between the initial velocity of the reaction and enzyme concentration, the rate of DNA methylation catalysed by M.HpyAVIB was determined. Duplex 1 (Table S1) was used as a substrate (with one 5′-CCTC-3′ site). Different concentrations of M.HpyAVIB (30–210 nm) were added to reaction mixtures containing duplex 1 (3 μm) and AdoMet (2.0 μm) and incubated at 37 °C. When the initial velocities were plotted against increasing enzyme concentrations, a linear relationship was obtained (data not shown). Further analysis of initial velocity versus DNA concentration or AdoMet concentration gave a Km (DNA) of 0.94 ± 0.06 μm and a Km (AdoMet) of 1.00 ± 0.05 μm (Table 1).

Table 1.   Kinetic parameters for M.HpyAVIB and in vitro translated M.HpyAVIB C5 cytosine MTase.
 M.HpyAVIBIn vitro translated M.HpyAVIB
inline image0.94 ± 0.06 μm0.45 ± 0.05 μm
inline image1.00 ± 0.05 μm0.65 ± 0.06 μm
kcat0.60 ± 0.06 min−10.55 ± 0.05 min−1
kcat/inline image1.0 ± 0.05 × 104 m−1·s−11.2 ± 0.05 × 106 m−1·s−1

Effect of divalent cations on the methylation activity of M.HpyAVIB

M.HpyAVIB has a putative metal-binding motif P114E115(X)20E136XK138, where the two acidic residues are important for the metal ion binding [22]. Divalent cations are known to affect the catalytic properties of a few MTase. For example, magnesium is an absolute requirement for methylation activity of M.EcoP15I, a member of the Type III RM systems [23]. It has been shown that metal ions act as a stimulator for M.Eco571 and M.BcgI activity [24,25]. We found that divalent cations stimulated the methylation activity of M.HpyAVIB. Calcium and nickel had the maximum effect on methylation because they enhanced the activity two-fold. When EGTA and EDTA were used in the presence of calcium and nickel, respectively, no enhancement in methylation activity was observed, further supporting the observation that metal ions can enhance DNA methylation by M.HpyAVIB (Fig. 1C). NaCl was used to maintain the same ionic strength in all the reactions.

Purification and characterization of mutant M.HpyAVIB proteins AdoMet-binding motif (F9N), catalytic motif (C82W) and metal-binding motif (P114L-E115Q)

All DNA MTase have conserved motifs for AdoMet binding (FXGXG) and catalysis (PCQ/DPPY/SPPY) [1]. Several studies performing mutational studies on amino acids in these motifs have been reported [26–28], revealing their role in binding and catalysis. Site-directed mutagenesis was performed on M.HpyAVIB to produce three mutant proteins: F9N (AdoMet binding motif), C82W (catalytic motif) and P114L-E115Q (metal binding motif). All three mutant proteins were purified to near homogeneity (data not shown). The methylation activity of F9N, C82W and P114L-E115Q mutant proteins was determined as a function of increasing enzyme concentration. It was found that both F9N (AdoMet binding motif) and C82W (catalytic motif) mutants were catalytically inactive compared to wild-type M.HpyAVIB (Fig. 2A). By contrast, in the presence of metal ions, the active P114L-E115Q mutant was unable to show any enhancement in methylation compared to the wild-type (Fig. 2B).

Figure 2.

 Characterization of M.HpyAVIB C82W, P114L and F9N mutant MTases. (A) Effect of enzyme concentration. Increasing concentrations of wild-type or mutant M.HpyAVIB (15–120 nm) incubated with 1 μm duplex 1 and 2.0 μm AdoMet in the presence of 10 mm Tris–HCl (pH 8.0), 100 mm NaCl, 5 mmβ-mercaptoethanol at 37 °C for 15 min. Reactions were stopped and analyzed as described in the Experimental procedures. •, Wild type; ♦, P114L-E115Q; ▲, C82W; □, F9N. (B) Comparison between the effect of Ca2+ (2 mm) on the methylation activity of M.HpyAVIB and the P114L-E115Q mutant. Methylation activity of wild-type M.HpyAVIB in the absence of calcium was taken as 100%.

hpyAVIBM undergoes random mutation when M.HpyAVIB is over-expressed in BL21(DE3) strain

M.HpyAVIB was purified for a second time from E. coli BL21(DE3) cells expressing the MTase. When the methylation activity of this purified MTase was measured, divalent metal ions inhibited the rate of methylation (Fig. 1D) in contrast to the earlier observation (Fig. 1C). One reason for the change in the property of the MTase could be the occurrence of mutations in the hpyAVIBM gene. To check this possibility, pET28a-hpyAVIBM plasmid DNA was isolated separately from an E. coli BL21(DE3) culture over-expressing M.HpyAVIB [after 4 h of induction with 1.0 mm isopropyl thio-β-d-galactoside (IPTG)] and an un-induced E. coli BL21(DE3) culture. hpyAVIBM was sequenced from the plasmids and, interestingly, mutations were observed in hpyAVIBM sequenced from plasmid isolated from the E. coli BL21(DE3) culture over-expressing M.HpyAVIB (Fig. S2A–E).

In further experiments, plasmid DNA was isolated after 1–4 h of induction, and mutations were observed only after 3 h (Fig. S2B,C). Both transition (Fig. S2D) and transversion (Fig. S2C,E) mutations were observed at the DNA level, which in turn resulted in silent mutation S to S (Fig. S2D) and missense mutations S to R and I to M (Fig. S2C,D). Figure S2A–E clearly shows that E. coli BL21(DE3) cultures over-expressing M.HpyAVIB had a mixture of pET28a-hpyAVIBM plasmids containing random mutations. Therefore, the M.HpyAVIB protein obtained from this culture had a mix of wild-type and M.HpyAVIB variants. As a control M.HpyAVIA (HP0050), an N6 adenine MTase from H. pylori strain 26695 and C82W or F9N M.HpyAVIB mutant in BL21(DE3) were over-expressed under the same conditions and the plasmids were sequenced. It was observed that over-expression of the M.HpyAVIA MTase or the F9N or C82W mutants of M.HpyAVIB did not induce mutations. Only the expression of wild-type M.HpyAVIB induced mutations and produced a ‘purified’ mixture of M.HpyAVIB proteins. This protein mixture was termed the ‘M.HpyAVIB variant’.

M.HpyAVIB variant has relaxed specificity

The M.HpyAVIB variant was checked for its sequence specificity. Different 26 mer duplex substrates (Table S1) with either CCTC (duplex 1), CTCC (duplex 2), CGCC (duplex 3), CCCC (duplex 4), CACC (duplex 5) or CTCA (duplex 6) were used in the methylation assay to check the specificity of M.HpyAVIB variant. It was found that M.HpyAVIB variant not only recognized and methylated CCTC, but also CNCC (where N is A, T, G or C), although not CTCA (Fig. 3). It was recently shown for E. coli Dam that a few amino acid changes in the MTase can lead to a change in specificity [29]. An in vitro methylation assay was then performed with the M.HpyAVIB variant (duplexes 1–5) (Table 2). Duplex 1 with a CCTC site was the preferred substrate with a Km of 0.4 μm. In addition, the kcat (turnover number) was calculated for different DNA substrates and wild-type MTase had a higher turnover (Table 1) with duplex 1 compared to the M.HpyAVIB variant (Table 2).

Figure 3.

 Methylation activity of M.HpyAVIB as a function of increasing concentrations of different 26 mer duplex DNA. Methylation assays were carried out in reactions containing [3H] AdoMet (2.0 μm) and increasing concentrations of 26 mer duplex DNA (0.25–3.5 μm) in standard reaction buffer at 37 °C. M.HpyAVIB (50 nm) was added to start the reaction. inline image, CCTC; ■, CTCC; ▲, CGCC; ▼, CCCC; ♦, CACC; •, CTCA.

Table 2.   Kinetic parameters on different DNA substrates for the M.HpyAVIB C5 cytosine MTase variant.
DuplexKMkcat (min−1)kcat/inline image (m−1·s−1)
10.40 ± 0.050.34 ± 0.051.42 ± 0.05 × 104
20.58 ± 0.080.18 ± 0.030.52 ± 0.08 × 104
30.52 ± 0.060.14 ± 0.030.44 ± 0.05 × 104
40.55 ± 0.070.18 ± 0.020.54 ± 0.05 × 104
50.45 ± 0.040.19 ± 0.030.70 ± 0.04 × 104

M.HpyAVIB variant methylates third cytosine in 5′-CTCC-3′

Wild-type M.HpyAVIB methylates the first cytosine in the recognition sequence 5′-CCTC-3′ [16,18]. Duplex 2 (Table S1) was used as a substrate to identify the target cytosine in 5′-CTCC-3′ methylated by the M.HpyAVIB variant. Duplex 2 contains a recognition sequence (CTCC) for the M.HpyAVIB variant with overlapping M.MspI (CCGG) and M.SssI (CG) sites as CTCCGG. M.MspI will methylate the second cytosine in the CTCC sequence (CTCCGG) and M.SssI will methylate the third cytosine in the CTCC sequence (CTCCGG).

Duplex 2 was either methylated by the M.HpyAVIB variant or by the M.MspI in the presence of radiolabelled AdoMet separately for 5 h and then the amount of label incorporated was measured. In a second experiment, duplex 2 was methylated by the M.HpyAVIB variant or by M.MspI, although, in the presence unlabelled AdoMet and after methylation, the enzyme was heat inactivated and duplex DNA purified by using QIAquick nucleotide removal kit (Qiagen, Valencia, CA, USA). Purified duplex DNA was then subjected to a second round of methylation in the presence of labelled AdoMet. In this second round of methylation, duplex methylated by mutant M.HpyAVIB was subjected to methylation by M.MspI and vice versa. Methylation by either of the MTase had no effect on the methylation by the second MTase, suggesting that second cytosine in CTCC was not the target cytosine for the M.HpyAVIB variant (Fig. 4A). When a similar experiment was performed using the M.SssI MTase, it was observed that methylation by one MTase had a significant inhibitory effect on the methylation by the second MTase (Fig. 4B). To confirm this, duplex 7 was used as a substrate for methylation by the M.HpyAVIB variant. Duplex 7 contains a CTCmC site where mC is C5 methyl cytosine. The M.HpyAVIB variant was unable to methylate duplex 7, indicating that third cytosine in CTCC was the target base for the M.HpyAVIB variant (Fig. S3).

Figure 4.

 Determination of target cytosine in CTCC. Competitive methylation assay. (A) Between M.HpyAVIB variant and M.MspI. (B) Between M.HpyAVIB variant and M.SspI. Duplex 7 was used as a substrate for the methylation reaction. M.HpyAVIB: reaction in the presence of [3H]AdoMet and M.HpyAVIB variant; M.MspI: reaction in the presence of [3H]AdoMet and M.MspI; M.HpyAVIB-M.MspI: reaction in the presence of unlabelled AdoMet and M.HpyAVIB and then with [3H]AdoMet and M.MspI; M.MspI-M.HpyAVIB: reaction in the presence of unlabelled AdoMet and M.MspI and then with [3H]AdoMet and M.HpyAVIB variant; M.SssI: reaction in the presence of [3H]AdoMet and M.SssI; M.HpyAVIB-M.SssI: reaction in the presence of cold AdoMet and M.HpyAVIB variant and then with [3H]AdoMet and M.SssI; M.SssI-M.HpyAVIB: reaction in the presence of unlabelled AdoMet and M.SssI and then with [3H]AdoMet and M.HpyAVIB variant. Recognition site of M.HpyAVIB variant is underlined and M.MspI and M.SssI recognition sites are shown in bold.

To further confirm that the third cytosine in CTCC was the target base for the M.HpyAVIB variant, a 250-bp DNA fragment with CCTC, CTCC and CCCC sites was used as a substrate for methylation by M.HpyAVIB variants and then subjected to bisulfite treatment as explained in the Experimental procedures. In most of the sodium bisulfite-treated plasmid samples, the first cytosine in CCTC, the third cytosine in CTCC and the fourth cytosine in CCCC showed protection from bisulfite treatment (Fig. S4A–C). In addition, it was the first cytosine in CCTC that showed protection from bisulfite treatment (Fig. S4C), whereas, in untreated DNA fragments, protection from bisulfite treatment was not observed (Fig. S4A). This clearly suggests that the M.HpyAVIB variant is a mixture of wild-type and mutant MTases, in which the wild-type enzyme methylates the first cytosine in 5′-CCTC-3′ and the variant is able to methylate the third cytosine in 5′-CTCC-3′.

Mutations alter the biochemical properties of M.HpyAVIB

Escherichia coli BL21(DE3) cells were used to over-express M.HpyAVIB protein and, after 4 h of induction by IPTG, a small aliquot of cells was plated on LB-kanamycin plates and incubated at 37 °C for 12 h. Next, randomly a single colony was picked from the plate, used to inoculated in 100 mL of LB supplemented with kanamycin and expression was induced by IPTG. The protein was purified and a small aliquot of E. coli BL21 (DE3) cells from the same induced culture was again plated on LB kanamycin plates. A single colony was picked from the plate and inoculated in 100 mL of LB, expression was induced, and the protein was purified. From four such different cultures, M.HpyAVIB was purified and biochemically characterized. All four protein preparations were catalytically active (data not shown). It was found that the M.HpyAVIB obtained from the four different preparations was different based on kinetic properties. M.HpyAVIB from cultures I and IV was able to methylate CCTC, whereas M.HpyAVIB from cultures II, III and IV showed relaxed specificity because it was not only able to methylate CCTC, but also CTCC and CCCC also (Table 3). Furthermore, M.HpyAVIB from culture III was able to methylate uracil when the target cytosine was replaced with uracil in duplex 8 (Table S1). It is worth noting that M.HhaI, a C5 cytosine Mtase, can methylate when the target cytosine was replaced with uracil [30]. When methylation assays were carried out in the presence of divalent cations such as nickel, it was observed that these metal ions enhanced the activity of M.HpyAVIB purified from culture I, whereas they had an inhibitory effect on M.HpyAVIB purified from cultures II and III (Table 3 and Fig. S5).

Table 3.   Comparison between M.HpyAVIB MTase purified from different cultures.
Culture numberEffect of metal on enzyme activitySpecificity of M.HpyAVIB variantMutations observed in variants
IActivationCCTC/CCCCC21Y, E23K, E63D, G79R
IIInhibitionCCTC/CTCC/CCCCT62A, K73Q, K74T
IIIInhibitionCCTC/CNCC/CNCUI53T, L141F, N142K, E135G
IVNo effectCCTC/CNCCG79R, E136G

Peptide fingerprint maps were obtained for each of the purified M.HpyAVIB from all four cultures. A peptide fingerprint map of M.HpyAVIB protein was obtained by digesting purified M.HpyAVIB protein with trypsin or chymotrypsin and subjecting it to MALDI-MS analysis. Peptide maps for purified M.HpyAVIB from cultures I, II and IV were compared, and it was found that 95% of peptides were similar in all proteins. However, there were three peptides (i.e. peptide a, which corresponds to amino acids 131–138, peptide b, which corresponds to amino acids 157–164, and peptide c, which corresponds to amino acids 36–46) present in M.HpyAVIB from cultures I and IV and absent in II (Fig. 5). The identity of the peptides was further confirmed by MS/MS (data not shown). Similar differences were observed when the peptide fingerprint map was obtained by digesting the proteins with chymotrypsin (Fig. S6). The differences between the peptide fingerprint maps of M.HpyAVIB from different cultures highlight the differences in amino acid sequence of these variants of M.HpyAVIB. As a control, the C82W mutant of M.HpyAVIB was used and no differences were observed in the peptide fingerprint maps.

Figure 5.

 Comparative peptide fingerprint map of M.HpyAVIB from cultures (A) I, (B) II and (C) IV. In total, 500 ng of purified protein was digested with trypsin (1 mg·mL−1) for 3 h at 37 °C. A peptide fingerprint map was obtained as described in the Experimental procedures.

When CD spectra were recorded for the M.HpyAVIB variants from different preparations, no significant differences were observed. However, when spectra were recorded in the presence of increasing concentrations of nickel, significant differences were observed between purified M.HpyAVIB from cultures II and IV (Fig. 6A,B). M.HpyAVIB purified from culture II showed a 28% decrease in helical content of the protein, whereas this decrease was absent in M.HpyAVIB from culture IV. M.HpyAVIA (HP0050) N6 adenine MTase was used as a control and did not exhibit any significant alteration in secondary structure in the presence of metal ions (Fig. 6C). The interesting differences observed in the CD spectra of purified M.HpyAVIB from cultures II and IV in the presence of metal ions could explain the differential effects of metal ions on the methylation activity of M.HpyAVIB from cultures II and IV. The pET28a-hpyAVIBM plasmid was also isolated from each culture and sequenced. A number of mutations were observed of which several were located in the target recognition domain (Table 3) and could be responsible for the change in the DNA specificity. It has been shown previously that in vitro mutagenesis coupled with DNA shuffling can result in relaxed specificity in SinI DNA-MTase [31].

Figure 6.

 Conformational changes in M.HpyAVIB variants in the presence of increasing concentrations of Ni2+. (A) CD spectra of M.HpyAVIB from culture I. (B) CD spectra of M.HpyAVIB from culture IV. (C) CD spectra of M.HpyAVIA. •, CD spectra in the absence of metal ion; CD spectra in the absence of metal ion at: ○, 1 mm; ▼, 2 mm; Δ, 3 mm.

Characterization of in vitro translated M.HpyAVIB

DNA sequencing of hpyAVIBM (Fig. S2) from plasmid pET28a-hpyAVIBM and peptide mapping of M.HpyAVIB variants (Figs 5 and S6) clearly showed that M.HpyAVIB purified from E. coli BL21(DE3) cells was a mixture of wild-type M.HpyAVIB and its variant proteins. To obtain M.HpyAVIB in a pure form without any mutations, pET28a-hpyAVIBM plasmid was used as a template and M.HpyAVIB protein was synthesized using the PURExpress™In vitro protein synthesis kit (New England Biolabs, Beverly, MA, USA) in accordance with the manufacturer’s instructions (Fig. 7). M.HpyAVIB protein was purified and kinetic parameters were determined (Table 1). It was observed that in vitro translated protein recognized duplex 1 with a recognition sequence CCTC and methylated the first cytosine only (data not shown). Duplex 1 was used as substrate for the methylation assay. In vitro translated M.HpyAVIB had a two-fold higher inline image and inline image and 100-fold better kcat/inline image compared to M.HpyAVIB purified from E. coli BL21 (DE3) cells (Table 1). In vitro translated protein did not have mutations compared to M.HpyAVIB purified from E. coli BL21(DE3) cells. It is therefore likely that M.HpyAVIB purified from E. coli BL21(DE3) cells is catalytically less active as a result of random mutations than in vitro translated M.HpyAVIB.

Figure 7.

In vitro translation of M.HpyAVIB. Lanes 1 and 2, purified M.HpyAVIB; lane 3, reaction without plasmid; lane 4, protein marker; lane 5, reaction with pET28a plasmid; lane 6, reaction with pET28a- hpyAVIBM.

hpyAVIBM exhibits high allelic variation between H. pylori strains

Helicobacter pylori exhibits a very high mutation rate compared to other bacterial species resulting in high genetic diversity [19,20]. In the present study, mutations in hpyAVIBM result in an alteration in the kinetic properties of the MTase. Thus, it was important to determine whether hpyAVIBM exhibits any allelic variation in H. pylori strains. hpyAVIBM homologues were amplified by PCR from different H. pylori strains isolated from patients with a diagnosis of duodenal ulcer (DU) on the basis of endoscopic examination of the stomach and duodenum, as well as adult healthy volunteers of both sexes who had no gastritis or dyspeptic syndromes. The sequence of hpyAVIBM homologues from different strains (San 51, PG 12, 516, PG 93, San 10 and I157) showed variations (Data S1). Compared to H. pylori strain 26695, M.HpyAVIB in strains PG 12, 516, PG 93, San 10 and I157 has 25, 20, 23, 17 and 32 amino acid differences, respectively (Data S1B). Interestingly, hpyAVIBM in strain San 51 has a 2-bp deletion, and thus codes for a truncated MTase (Data S1B). The variations observed in M.HpyAVIB between H. pylori strains are very much similar to the mutations observed in the MTase variants purified from E. coli described above. Thus, the variations that were observed in the hpyAVIBM homologues could result in an alteration of the kinetic properties of the MTase homologues in vivo.

Discussion

The presence of mutator cells is not unusual in a natural bacterial population. Most of the time, the mutator cells are characterized by a defective mismatch repair pathway [32–34]. Mutator cells are very common in pathogenic bacteria such as H. pylori, Neisseria sp. and E. coli, and play a critical role in enhancing the adaptive capacities of the pathogens in the host. Stress increases the percentage of mutator cells in the population, thus providing an opportunity to overcome the environmental challenge [33–36].

In the present study, we show that over-expression of a C5 cytosine MTase from H. pylori, M.HpyAVIB, creates a mutator phenotype in E. coli BL21(DE3) cells. A number of studies have suggested a link between cytosine MTase and enhanced mutation rates because methylated cytosine acts as a hotspot for mutation (5mC to T mutation) [9,13,37]. It has also been shown that deoxycytosine MTase enzyme activity causes mutagenesis in vitro either by deamination of cytosine to uracil in the absence of AdoMet or, indirectly, through spontaneous deamination of C5 methyl-cytosine to thymine [13]. For the expression of M.HpyAVIB, we have observed enhanced mutation rates but C5 methyl-cytosine to thymine mutations were not observed. In the present study, we report a possible alternative mechanism for the enhanced mutation rates in E. coli upon over-expression of M.HpyAVIB. The data reported indicate the creation of an epigenetic circuit in the E. coli cell upon induction of M.HpyAVIB resulting in changes to the transcriptome. It is possible that over-expression of M.HpyAVIB results in an alteration of the repair pathways, leading to transitions and transversions. Most interestingly, the MTase itself has mutated in various ways to display different catalytic properties. M.HpyAVIB variants have different specificities and respond differentially in the presence of metal ions (Table 3). A significant point worthy of note is that H. pylori lacks functional mismatch repair [14,15]. The overall mutation rates are high in H. pylori compared to other bacteria. hpyAVIBM and other genes are subjected to mutations continuously and mutations coupled with selection could be a strong force in the evolution of novel genes with new properties. Sequencing of hpyAVIBM homologues from different H. pylori strains has revealed a high degree of variation. The results obtained in the present study clearly demonstrate that variants of M.HpyAVIB show different kinetic properties (Table 3). Creation of diversity could be critical for an organism such as H. pylori, which has a limited host range to survive and cope with an ever changing micro-environment in the host.

Experimental procedures

Strains and plasmids

Helicobacter pylori 26695 strain (cagA+iceA1 vacAs1m1) genomic DNA was obtained as a gift from New England Biolabs. E. coli strain DH5α [F′end A1 hsd R17 (inline imageinline image) glnV44 thi1 recA1 gyrA (NalR) relA1 Δ (lacIZYA –argF) U169 deoR (Φ80dlacΔ (lacZ)M15)] was used as a host for preparation of plasmid DNA. E. coli BL21 (DE3) pLysS-FompT hsdSB (inline imageinline image) gal (dcm)(lon) (DE3) pLysS cells were used for expression of wild-type and mutant M.HpyAVIB proteins. H. pylori cultures were grown on petri plates containing brain heart infusion agar (Difco, Franklin Lakes, NJ, USA) with horse blood or serum, isovitalex, and antibiotics [vancomycin (6 μL·mL−1), trimethoprim (8 μL·mL−1) and polymyxine B sulfate (2.5 U·mL−1)].

DNA manipulation and analysis

Chromosomal DNA from H. pylori was prepared from confluent growth on brain heart infusion agar plate cultures by the cetyltrimethylammonium bromide extraction method [38]. PCR for detection of the hpyAVIBM allele was carried out by using the appropriate primers (primers 1 and 2; Table S2). Positive and negative controls were included in each assay. PCR products were sequenced. PCR products were ligated into vector pTZ57R/T (Fermentas, Glen Burnie, MD, USA) in accordance with the manufacturer’s instructions.

PCR amplification and cloning of hpyAVIBM gene of Helicobacter pylori 26695

The 1064-bp long hpyAVIBM gene was amplified from genomic DNA of H. pylori 26695 strain by PCR with Pfu polymerase using primers 1 and 2 (Table S2). The primers were designed with the help of the annotated complete genome sequence of H. pylori 26695, considering the putative gene sequence of hpyAVIBM, obtained from TIGR. The amplified PCR fragment was ligated into the EcoRV site of pETBlue-1 (Novagen, Madison, WI, USA) and then inserted into the bacterial expression vector pET28a at the BamHI and XhoI sites.

Over-expression of M.HpyAVIB C5 cytosine MTase of Helicobacter pylori

Escherichia coli strain BL21 (DE3) cells were transformed with the pET28a-hpyAVIBM DNA construct. Individual colonies obtained after transformation were used to inoculated LB broth containing 50 μg·mL−1 kanamycin and grown overnight at 37 °C. In total, 1% of this primary inoculum was then used for reinoculation and grown until A600 of 0.6 was reached. M.HpyAVIB protein production was induced by the addition of 1 mm IPTG. After 1 h of incubation at 30 °C, the culture was cooled on ice and an approximately equal number of bacterial cells from un-induced and induced culture were harvested by centrifugation at 2300 g for 10 min. The induction of the protein was checked by 10% SDS/PAGE of crude cell extract obtained by sonication in SDS/PAGE buffer containing dye. As a control, protein expression in E. coli BL21 (DE3) cells and the same cells containing pET28a vector only was examined.

Purification of wild-type and mutant M.HpyAVIB C5 cytosine MTases

Escherichia coli BL21 (DE3) cells harbouring pET28a- hpyAVIBM constructs were grown in 600 mL of LB broth containing 50 μg·mL−1 kanamycin until A600 of 0.6 was reached and M.HpyAVIB protein expression was induced by the addition of IPTG to a final concentration of 1 mm, at 30 °C. After 1–4 h of induction at 30 °C, the culture was cooled on ice and cells were harvested by centrifugation at 5900 g for 30 min at 4 °C.

All purification steps were carried out at 4 °C. The cell pellet was resuspended in extraction buffer (10 mm Tris–HCl, pH 8.0, 0.05% Triton-X, 100 mm NaCl and 50 mm imidazole) and lysed by sonication. The cell lysate was centrifuged at 5900 g for 1 h at 4°C. Supernatant was collected and M.HpyAVIB protein was loaded onto a nickel-nitrilotriacetic acid column that was previously equilibrated with the above mentioned buffer. The enzyme was eluted by using 10 mL of 10 mm Tris–HCl (pH 8.0) containing 100 mm NaCl and 200 mm imidazole. The eluted enzyme was dialysed overnight at 4 °C against 10 mm Tris–HCl (pH 8.0) containing 100 mm NaCl, 10 mmβ mercaptoethanol and 30% glycerol. The purity of the protein preparation was judged by SDS/PAGE with Coomassie brilliant blue staining [39]. Protein concentration was estimated by the method of Bradford using BSA as standard [40].

Site-directed mutagenesis

Site-directed mutagenesis was performed using a PCR-based technique to replace the required amino acids [41]. Mutations were introduced into the hpyAVIBM gene by using the two-stage megaprimer PCR method. PCR reactions were carried out with Phusion DNA polymerase (Finnzymes Oy, Espoo, Finland). For each substitution, a mutagenic primer and an appropriate second primer was used. In the first round of PCR, oligonucleotide primers (Table S2) and pET28a-hpyAVIBM DNA were used to amplify a DNA fragment, which was used as a megaprimer in the second round of PCR. The full-length PCR product was obtained in the second-round PCR by extension of the megaprimer. The PCR product thus obtained was purified, digested with DpnI restriction enzyme to cleave the methylated template DNA, transformed into E. coli DH5α strain and plated on LB agar media containing kanamycin (50 μg·mL−1). The mutagenic primers were designed in such a way to change the respective amino acids and to create a Type II restriction enzyme site. Hence, the resultant plasmids could be screened easily. The resultant plasmids were used for expression and purification of mutant M.HpyAVIB proteins. The amino acid (shown in bold) in the AdoMet binding motif FXGXG was replaced by using primers 3 and 6 (primer 3 was mutagenic; Table S2). By substituting F with N, it was possible to introduce convenient restriction site (DraI), thus allowing screening of F9N mutants. Similarly, site-directed mutagenesis was performed to replace one amino acid (shown in bold) in the catalytic site PCQ by using primers 4 and 6 (primer 4 was mutagenic). By substituting C with W, a NcoI site was created and this property was used for the screening of C82W mutants. The metal-binding motif P114E115(X)20E136XK138 was mutated by using primers 5 and 6 (primer 5 was mutagenic). By replacing P114 by L and E115 by Q, a PstI site was created. The mutants were confirmed by restriction digestions and DNA sequencing.

Methylation activity

In vitro methylation

All methylation assays monitored incorporation of tritiated methyl groups in to DNA by using a modified ion-exchange filter binding assay [42]. Methylation assays were carried out in a reaction mixture (20 μL) containing 26 mer duplex (duplex 1–8; Table S1) that harbours a single recognition sequence in the centre, [3H]AdoMet (specific activity 66 Ci·mmol−1) and purified protein in the reaction buffer (10 mm Tris–HCl, pH 8.0, 100 mm NaCl, 5 mmβ-mercaptoethanol). After incubation at 37 °C for 20 min, reactions were stopped by snap-freezing in liquid nitrogen. Background counts were measured at zero-time incubation, the incubation in the absence of enzyme was subtracted, and the data were analyzed. All methylation experiments were carried out at least in triplicate and the results averaged. Standard deviations of the average methylation rates were below 10%.

Competitive methylation assay

To investigate methylation by M.HpyAVIB, 300 pmol of a 26 mer duplex containing one CTCC site with an overlapping M.SssI and M.MspI site (duplex 2; Table S1) was incubated with 15 μm AdoMet with M.HpyAVIB or M.SssI or M.MspI in their respective reaction buffer and incubated for 5 h at 37 °C separately followed by protein inactivation at 95 °C for 20 min. DNA was purified and methylated duplex 2 was subjected to second round of methylation in the presence of [3H]AdoMet. When the methylation was carried out for 5 h, fresh AdoMet was added to the reaction mix every 1 h.

Bisulfite sequencing

A 250-bp DNA fragment was used as a substrate for the M.HpyAVIB methylation assay. After methylation, DNA was purified and treated with sodium bisulfite [43] purified on a spin-column (Qiagen). The 250-bp DNA was amplified by PCR using primers. Untreated DNA, amplified by the same pair of primers, served as a control. The PCR products were gel-purified and ligated into the EcoRV site of pETBlue1 vector (Novagen) and, from each experiment, plasmids from five individual clones were sequenced and the conversion of cytosines to thymines monitored. Sodium bisulfite converts unmodified cytosine to uracil and these are amplified as thymidine in PCR, whereas methylated cytosines are protected from this modification. Sodium bisulfite-treated DNA sequences were compared with control sequences to identify the sites containing C5 methyl cytosine.

MS

MS data were acquired by MALDI-MS (Ultraflex TOF/TOF; Bruker Daltonics, Ettlingen, Germany), with a N2 laser emitting radiation at a wavelength of 337 nm. All the data were recorded in reflectron positive ion mode. The matrices were α-cyano 4-hydroxy cinnamic acid (α-C) or 2,5-dihydroxy benzoic acid. Each mass spectrum was the sum of data acquisitions from 200 to 400 laser shots. The tandem mass spectrum (MS/MS) was acquired using the LIFT method, a modified form of post-source decay. Each MS/MS spectrum was the sum of data acquisitions from 500 to 700 laser shots; 200–300 shots for precursor ion isolation and 300–400 shots for recording product (fragment) ion spectrum. A high-resolution timed ion selector aids in precursor ion isolation [44]. The mass window (ΔM) for precursor ion isolation was set to ± 10 Da. The data were processed using flexanalysis, version 2.4 (Bruker Daltonics).

Determination of kinetic parameters

Kinetic studies were performed using 26 mer duplex (duplex 1–5; Table S1). Methylation assays were carried out as described earlier in a series of similar reactions containing M.HpyAVIB (50 nm), [3H]AdoMet (2.0 μm) and 26 mer duplex (0.25–3.5 μm) (Table S1). The velocities were fitted into a one-site binding (hyperbola) equation:

image

All data were analyzed by nonlinear regression. A double reciprocal plot of initial velocity versus DNA (26 mer duplex) concentration allowed the determination of Km (DNA) and Vmaxkcat was calculated as ratio of Vmax/[E]. Similarly, initial velocity experiments were carried out by varying the concentration of [3H]AdoMet in the range 0.3–2.4 μm at the same time as keeping the DNA concentration fixed at 2 μm, and also keeping other reaction conditions identical. The double reciprocal plot of initial velocity versus AdoMet concentration allowed the determination of Km (AdoMet). The nonlinear regression analysis of initial velocity versus AdoMet concentration allowed the determination of Km (AdoMet). Data were analyzed by nonlinear regression analysis (curve fit) using graphpad prism, version 5 (GraphPad Software Inc., San Diego, CA, USA). All methylation experiments were carried out in triplicate and the results averaged. SDs of the average methylation rates were below 10%.

Miscellaneous

CD spectral analysis, dynamic light-scattering, molecular mass and oligomeric status determination were carried out as described previously [45].

Acknowledgements

R.K. thanks the CSIR for a Senior Research Fellowship. All members of the DNR laboratory are acknowledged for critically reading the manuscript and useful discussions. The work was aided by a grant from the Department of Biotechnology, Government of India to D.N.R. We thank the DBT mass spectrometry facility at the Indian Institute of Science for peptide fingerprint mapping. D.N.R. acknowledges the DST for a J.C. Bose Fellowship.

Ancillary

Advertisement