Specific targeting of cytosine methylation to DNA sequences in vivo

Abstract

Development of methods that will allow exogenous imposition of inheritable gene-specific methylation patterns has potential application in both therapeutics and in basic research. An ongoing approach is the use of targeted DNA methyltransferases, which consist of a fusion between gene-targeted zinc-finger proteins and prokaryotic DNA cytosine methyltransferases. These enzymes however have so far demonstrated significant and unacceptable levels of non-targeted methylation. We now report the development of second-generation targeted methyltransferase enzymes comprising enhanced zinc-finger arrays coupled to methyltransferase mutants that are functionally dominated by their zinc-finger component. Both in vitro plasmid methylation studies and a novel bacterial assay reveal a high degree of target-specific methylation by these enzymes. Furthermore, we demonstrate for the first time transient expression of targeted cytosine methyltransferase in mammalian cells resulting in the specific methylation of a chromosomal locus. Importantly, the resultant methylation pattern is inherited through successive cell divisions.

Figure 1

(A) Methylation sensitive restriction analysis of expression plasmids coding for wild-type and mutant targeted methyltransferases. Expression plasmids coding for 4aWT, 4aFH and 4aNull HpaII based enzymes and 4aWT, 4aQG and 4aNull HhaI based enzymes were isolated from E.coli and subjected to restriction with methylation sensitive and/or insensitive isoschizomer restriction enzymes (left and right panels). The middle panel shows methylation-sensitive restriction digestion of higher copy number GST fusion expression plasmids coding for HpaII wild-type and FH mutants and as fusions with the 4a zinc-finger gene. Digestion products are shown. U, uncut; Hp, HpaII; Ms, MspI; Hh, HhaI restriction enzymes. (B) SDS–PAGE analysis of purified recombinant zinc-finger–Mtase fusion proteins, used for in vitro binding and methylation specificity studies.

Figure 2

(A) EMSA analysis of purified recombinant four zinc-finger–HpaII (F³⁵H) mutant Mtase fusion proteins (the 4a, 4b and 4c zinc-finger components of each protein are described in the text), compared with recombinant non-zinc-finger fusion M.HpaII wild-type and FH proteins. Recombinant protein was titrated against either a target site or non-target site containing probe, at 0.1, 0.3, 0.5, 1, 3, 5, 10, 30 and 50 fmol. Lane C is probe only. The target sites are given under the relevant panels. (B) Binding of M.HpaII wild-type protein to HpaII site DNA at equivalent concentration range to (A) is shown in the left-hand panel. Non-zinc-finger fusion M.HpaII wild-type and FH proteins were also used at 50-fold higher levels, so that some comparative evidence of binding by the HpaII FH protein to HpaII site DNA could be seen (middle and right-hand panels). (The probe used was Tb/c). (C) Methylation analysis of 4aFH versus M.HpaII protein for target and non-target site containing DNA fragments. Increasing amounts of each protein (25, 50, 100, 150, 250 and 350 fmol) were incubated with DNA fragments (1.0 μg) containing just an HpaII site (M), or an HpaII site and flanking 4a zinc-finger recognition site (ZfM). Following incubation, the reaction was stopped and DNA subjected to restriction with R.HpaII and products resolved by electrophoresis. The intensity of the uncut band was measured for each lane (AIDA software, raytest, GmbH) and normalized against the end point intensity. These data are presented in graphical form in the right-hand panel.

Figure 3

(A) Schematic diagram of the basis for plasmid-based and in situ bacterial assays for targeted methylation. A detailed diagram of the target site (which is embedded in the target site vector) is shown. Following either incubation of plasmid with recombinant targeted methyltransferases or isolation from bacteria co-transformed with expression and target site vectors, plasmids were subsequently restricted with either HhaI or HpaII restriction enzyme and the digest products were analysed on an agarose gel. In the absence of any methylation, restriction occurs at all sites and a standard fragment pattern is generated, which includes fragments of x and y bp in size derived from cleavage at a, b and the target site, indicated by a filled lollipop. Where targeted methylation has occurred, restriction at that site is blocked, thus modulating the normal restriction pattern by the generation of a unique band of predicted length (designated the target band), of size (x + y) bp as indicated in the figure. As further confirmation, this ‘target fragment’ should restrict at the EcoRV restriction sites within the target region. (B) Plasmid-based methylation analysis of targeted and non-targeted HpaII wild-type and FH mutant proteins. Linearly increasing amounts of purified protein (20–100 fmol, in 10 fmol increments) were incubated with either target site or non-target site plasmids (1–2 μg). After HpaII cleavage, the restriction products were analysed by gel electrophoresis. Arrows denote the appearance (or expected location for wild-type enzyme panels) of a DNA fragment corresponding to 560 bp in size [(x + y) fragment in Figure 1B], denoting targeted methylation. M is a DNA marker mix (NEB, 100 bp + 1 kb ladders). Lane C represents DNA digested in the absence of Mtase incubation. Plasmids were additionally restricted with EcoRV to confirm the identity of the 560 bp fragment. A typical result is shown for the 4aFH protein (centre panel).

Figure 4

(A) Plasmid-based methylation analysis of targeted and non-targeted HpaII wild-type and FH mutant proteins for ‘large fragment’ target vector. Increasing amounts of purified 4aFH and 4cFH protein (as described in Figure 3) were incubated with target site or non-target site plasmids. After HpaII cleavage, the restriction products were analysed via gel electrophoresis. Arrows denote the appearance of a DNA fragment corresponding to 1060 bp in size, denoting targeted methylation. M is a DNA marker described previously. Bands derived from non-specific methylation are shown using brackets. (B) Bacterial-based in vivo analysis of targeted methylation by 4cFH protein. Details are described in Materials and Methods. The left-hand panel shows the HpaII restriction profile of the 4cNull and target vectors alone, and after co-isolation from bacteria, for comparative purposes. The right-hand panel shows the restriction profile of 4cFH after co-isolation with either target (Tb/c2) or non-target (Ta2) vector. The band indicative of targeted methylation is arrowed. Low molecular weight target vector associated products are shown using brackets. A higher molecular weight product, just above the target band, resistant to KpnI cleavage, is shown using asterisks. H, HpaII; V, HpaII + EcoRV; K, HpaII + KpnI restriction in all cases. (C) Bacterial-based in vivo analysis of targeted methylation by 6aFH protein. Experimental details are essentially as in (B) above. The 1060 bp fragment indicative of targeted methylation is arrowed in the figure. The right-hand panel shows the results of digestion of plasmids isolated from culture grown from 3-day-old bacterial colonies. The 1060 bp target band is arrowed. Higher molecular weight bands related to the targeted methylation event are shown using brackets.

Figure 5

Methylation-sensitive restriction analysis of plasmids isolated from bacteria transformed with 4aHhaI or 6aHhaI expression vectors and target vector [Ta2(HhaI)]. Experimental details are essentially as in Figure 4B. (A) Restriction analysis of 4aQ²³⁷HhaI mutant or null vectors and 4aHhaI-specific target vector isolated from co-transformed bacteria. The 940 bp band denoting targeted methylation for 4aQG is shown using an arrow. Evidence of non-specific methylation by the 4QV mutant is shown using asterisks. (B) Left-hand panel shows the results of the bacterial assay described above, for the 6aQG protein. The band denoting targeted methylation is indicated (white arrow). The non-specific band is shown using an asterisk. The right-hand panel shows the effects of a transient increase in 6aQG expression vector copy number on the intensity of the 940 bp band and the appearance of higher molecular weight, targeted methylation related bands that disappear upon EcoRV cleavage (small arrows). Low molecular weight non-specific bands are shown using bracket. H, HhaI; V, HhaI + EcoRV restriction, throughout this figure. M is marker mix, Kb is 1 kb ladder (NEB).

Figure 6

Analysis of genomic DNA from bacteria expressing targeted methyltransferase enzymes. (A) Restriction analysis of genomic DNA recovered from bacteria expressing various 4aHhaI enzyme derivatives (and containing the target site vector), described above the relevant lanes. The lane denoted ‘Con’ represents genomic DNA isolated from untransformed bacteria. U, uncut; C, cut with HhaI restriction enzyme. (Fragments seen in some ‘cut’ lanes are derived from plasmid carryover during genomic DNA preparation, shown using arrows. Input genomic DNA is shown using asterisks.) (B) Arbitrary primed PCR of the HhaI restricted genomic DNA described in (A). (C) Restriction analysis of genomic DNA recovered from bacteria expressing various 4aHpaII enzyme derivatives, described above the relevant lanes. pWt is plasmid DNA coding for the wild-type targeted enzyme. U, uncut, C, cut with HpaII. (D) Arbitrary primed PCR of HpaII restricted genomic DNA described in (C). M is 1 kb ladder (NEB), 100bp is 100 bp ladder (NEB).

Figure 7

(A) Schematic diagram describing mammalian vectors used for the expression of targeted Mtases. (B) Upper panel showing western blot confirming expression levels of 4aFH, 4bFH and 4aQC targeted Mtase enzymes from vector described in (A). Lower panel showing the results of RT–PCR of 4aZfFH RNA at 3, 6 and 12 days post-transfection with 4aZfFH targeted Mtase expression vector. β-Actin was used an internal control. PCR was performed for 28, 30, 32 and 34 cycles. (C) Bisulphite sequencing results for genomic target site DNA, 14 days post-transient transfection with targeted Mtase or control vector. The target site sequence is given at the top of the figure. HpaII sites are in boldface and the 12 bp zinc-finger recognition site is boxed. Below this is a schematic diagram showing target site and flanking CpG site distribution (represented by circles in sequencing data) spanning ∼160 bp. Large ticks indicate HpaII sites, small ticks indicate HhaI sites. Zinc-finger recognition sites for the 4aZf protein are represented by filled rectangles. The expression vector transiently transfected in each experiment is indicated next to each sequencing results block. Filled circles indicate observed methylation.