Figure 1.
Schematic depictions of sequences and nomenclature of modeled protein/DNA complexes.
(A) Sequences of zinc fingers fused to fragments of M.HhaI methyltransferase. Numbers in brackets correspond to the amino acid numbers. Black segments correspond to linker sequences. (B) The orientation of the zinc finger binding sites relative to the intended methylation target site (the circle). The orientations depicted are the ones that would position the indicated protein pairs over the targeted CpG site. (C) Molecular model of a particular NZ/CZ construct containing the indicated linkers. The base to be methylated is indicated in purple. Models of other complexes can be found in Figure S1.
Figure 2.
A schematic of the restriction enzyme protection assay for targeted methylation.
(A) A single plasmid, pDIMN8, encodes genes for both methyltransferase fragment-zinc finger fusion proteins, as well as two sites for assessing the degree of targeted methyltransferase activity. Expression of both protein fragments was induced in ER2267 cells and plasmid DNA was isolated. (B) Plasmid DNA was linearized by NcoI-HF digestion and incubated with FspI, an endonuclease whose activity is blocked by methylation. In the absence of methylation, the plasmid is digested twice by FspI and once by NcoI-HF as shown. (C) Methylation at one or both of the FspI containing sites creates unique digestion patterns as assessed by agarose gel electrophoresis. Unique bands are diagnostic of no methylation (∼4600 bp), methylation at site 1 (∼5210 bp), methylation at site 2 (∼5830 bp), or methylation at both sites (∼6580 bp). (D) A schematic of the functional methyltransferase at a target site. Zinc finger/DNA recognition mediates methyltransferase assembly. (E) This assembly is designed not to occur at the non-target control site, which lacks zinc finger binding sites.
Figure 3.
The effect of C-terminal truncation, linker lengths, and target site spacing on methyltransferase activity.
(A) A schematic of the protein fusions and target DNA sequences indicating the variability in linker length and DNA spacing tested. The linkers connecting the zinc fingers to the N- and C-terminal fragments were varied in 5 amino acid increments (from 0 to 15 amino acids), and combined iteratively. The bases separating the FspI site from the zinc finger binding sites were also varied (0,1,2,3 bases on each side). (B) Truncation of the C-terminus of the C-terminal fragment (indicated in units of amino acids) decreases off-target activity at the methyltransferase. In this experiment X = 3, Y = 1 and Z = 0. The nature of the DNA at site 1 and site 2 (whether a target or non-target site) is depicted at the bottom of the figure and graph. Constructs in which the C-terminus of M.HhaI was truncated by 6 amino acids were used to determine the effect of (C) linker length and (D) target site spacing on methyltransferase activity at the target site. The percent methylation at the target site are indicated in the graphs. All graphs show the mean and the error bar represents the standard deviation of the analysis of plasmid DNA from n ≥3 independent cultures.
Figure 4.
Molecular modeling explains how an increase in target spacing can reduced the required protein linker length.
(A) HS1 zinc fingers are bound to DNA with a target site spacing of Z = 0,1,2, or 3. Note that at Z = 3, the zinc finger is actually closer to the bisection point of the N-termini than at Z = 0. (B) A model demonstrating that a five amino acid linker is sufficient to connect zinc finger HS1 bound at Z = 3. In contrast, a longer amino acid linker is required to circumvent the DNA backbone at Z = 0. This model provides an explanation for the pattern of target site methylation observed in Figure 3D.
Figure 5.
The contribution of each zinc finger binding site toward observed, targeted DNA methylation.
Methylation was assessed as in Figure 2. In this experiment, the C-terminal fragment of M.HhaI is truncated by 6 amino acids, X = 3, Y = 1, and Z = 0. Methyltransferase activity was assessed with and without target sites present. Moving the target site from site 1 to site 2 did not have a large effect on activity. Target half sites (in which either the HS1 or HS2 binding sites were removed) allowed assessment of the contribution of each zinc finger on methylation activity at the target site. The sum of the methylation observed on each half site (43±5%) was less than methylation at the full target site (61±6%). The methylation observed with two distal half sites (<30%) was also less than that observed with the complete target site.
Figure 6.
M.SssI can be converted into a targeted heterodimeric methyltransferase.
(A) A schematic showing the sequence of the M.SssI fragments fused to zinc fingers via flexible linkers. (B) A restriction enzyme protection assay showing the split enzyme constructs possess a bias for methylation at the target site. Plasmids were isolated from strains grown under conditions that either repress or induce expression of the two fragments. Plasmid DNA was assayed for methylation as in Figure 2. The activity of these fusion heterodimers was attenuated by the indicated point mutations known to decrease enzyme activity in wild-type M.SssI.