Proximal recognition sites facilitate intrasite hopping by DNA adenine methyltransferase: mechanistic exploration of epigenetic gene regulation

Abstract

The methylation of adenine in palindromic 5'-GATC-3' sites by Escherichia coli Dam supports diverse roles, including the essential regulation of virulence genes in several human pathogens. As a result of a unique hopping mechanism, Dam methylates both strands of the same site prior to fully dissociating from the DNA, a process referred to as intrasite processivity. The application of a DpnI restriction endonuclease-based assay allowed the direct interrogation of this mechanism with a variety of DNA substrates. Intrasite processivity is disrupted when the DNA flanking a single GATC site is longer than 400 bp on either side. Interestingly, the introduction of a second GATC site within this flanking DNA reinstates intrasite methylation of both sites. Our results show that intrasite methylation occurs only when GATC sites are clustered, as is found in gene segments both known and postulated to undergo in vivo epigenetic regulation by Dam methylation. We propose a model for intrasite methylation in which Dam bound to flanking DNA is an obligate intermediate. Our results provide insights into how intrasite processivity, which appears to be context-dependent, may contribute to the diverse biological roles that are carried out by Dam.

FIGURE 1.

Known examples of methylation state-dependent epigenetically regulated operons. A, legend. B, operons in the “on” state (left panel) and in the “off” state (right panel).

FIGURE 2.

Schematic of intrasite and intersite processivity. A, schematic of the intrasite processivity of Dam (open triangle) where each adenine in the palindromic GATC site is methylated. Arrows represent how Dam must both switch strands and rotate 180°. B, intersite processivity experiment where enzyme encounters a piece of DNA and can modify sites on both strands before dissociation. See the Introduction and references for diversity of enzymes and modifications (–26). Sites to be modified are represented as closed rectangles, and modifications to the sites are represented as closed circles.

FIGURE 3.

Validation of DpnI assay to detect intrasite processivity. Comparison of tritium and DpnI assay, 200 nm DNA, 210 nm Dam, 30 μm S-adenosylmethionine, and 15 °C are shown. 100% conversion for the DpnI trace refers to complete methylation of substrate and subsequent DpnI digestion based on the assay described in supplemental Fig. 2. Tritium data are represented by black dots, and DpnI data for the 115-bp fragment are represented by inverted gray triangles. Error bars represent between 2 and 5 replicates, mean ± S.D. A, substrate 1B. Single turnover fit for tritium is a solid black line, and single turnover fit for the DpnI data is a dashed gray line. B, tritium data (closed circle) and DpnI data (inverted gray triangle) for substrate 1C. C, reaction scheme for total methylation of an unmethylated site. k₁ is the methylation rate constant from unmethylated to hemimethylated; k₂ is the methylation rate constant from hemimethylated to doubly methylated. D, rate constants k₁ = 0.10 min⁻¹ and k₂ = 0.053 min⁻¹ are used in Equation 3 (dashed gray line) and Equation 4 (solid gray line) to fit with the data. Also included is the single turnover fit from the tritium data (solid black line), and the profile for hemimethylated DNA (gray diamonds).

FIGURE 4.

Intrasite processivity is modulated by lengths of flanking DNA. The tritium data (closed circle) and DpnI data (inverted gray triangle) for substrates from Table 1. k₁ and k₂ (as described under “Results” and in the legend for Fig. 3, min⁻¹) are given for the single site sequential substrates: substrate 1A (A); substrate 1D (B); k₁ = 0,072, k₂ = 0.036, substrate 1E (C); k₁ = 0,025, k₂ = 0.014 substrate 2B (D); substrate 2C (E); substrate 2D (F), substrate 2E (G); and substrate 1LL (H). Although the kinetic scheme is too complicated to predict what the sequential methylation of each site would be for substrate 2E, the characteristic delay in the DpnI trace in comparison with the tritium trace is convincing enough to assume that the hemimethylated intermediate is present. For D–G, the DpnI data represent the accumulation of the 115-mer fragment, which was nearly identical to the 119-mer fragment (not shown). Error bars represent between 2 and 5 replicates, mean ± S.D.

FIGURE 5.

Competition experiment with nonspecific DNA. The tritium data (closed circle) and DpnI data (inverted gray triangle) for substrate 1B with a 500-bp piece of chase DNA included are shown. Single exponential fits for tritium and Dpn1 are a black line and a dashed gray line, respectively. k_chem for the reaction is 0.011 ± 0.001 min⁻¹. Error bars represent between 2 and 5 replicates, mean ± S.D.

FIGURE 6.

Intrasite processivity of Dam mutants. The tritium data (closed circle) and DpnI data (inverted gray triangle) for Dam mutants with a 60-bp substrate are shown. A, K139A; B, N132A; C, R116A; D, R95A; E, N126A. Only E is sequential. Error bars represent between 2 and 5 replicates, mean ± S.D.

FIGURE 7.

Potential models of intrasite processivity and its regulation. The direct mechanism (mechanism i, under “Discussion”) is depicted in gray. The indirect mechanism (mechanism ii, under “Discussion”), where intrasite processivity proceeds by an intermediate with flanking DNA, is shown in black. Notably, the translocation step in mechanism i represents a loss of intrasite processivity. For mechanism i, the observed rate of the reaction and the occurrence of intrasite processivity would not be predicted to change with increases of flanking DNA.

FIGURE 8.

How flanking DNA regulates intrasite processivity. A, E is Dam, S is hemimethylated DNA, and P is fully methylated DNA. Shown is a schematic depicting the possible outcomes following the initial methylation of a GATC site. Either the enzyme will undergo intrasite processivity (scheme 1), or the enzyme will leave the hemimethylated substrate (scheme 2). B, the type of substrate dictates which mechanism occurs from A (scheme 1 or scheme 2). In i, the enzyme stays associated with the DNA long enough to remethylate it. However, in ii, the enzyme leaves the DNA because it spends too much time on the nonspecific DNA away from its GATC site, forcing scheme 2. In iii, the second GATC site allows Dam to spend longer on the DNA, pushing the reaction toward scheme 1.