Targeted regulation of imprinted genes by synthetic zinc-finger transcription factors


Epigenetic control of transcription is essential for mammalian development and its deregulation causes human disease. For example, loss of proper imprinting control at the IGF2–H19 domain is a hallmark of cancer and Beckwith–Wiedemann syndrome, with no targeted therapeutic approaches available. To address this deficiency, we engineered zinc-finger transcription proteins (ZFPs) that specifically activate or repress the IGF2 and H19 genes in a domain-dependent manner. Importantly, we used these ZFPs successfully to reactivate the transcriptionally silent IGF2 and H19 alleles, thus overriding the natural mechanism of imprinting and validating an entirely novel avenue for ‘transcription therapy’ of human disease.


Epigenetic states organize cell memories that dictate when, where and how genes are expressed during development. While the acquisition and maintenance of epigenetic states are normally of high fidelity to progressively restrict developmental potentials in lineage-specific manners, epigenetic lesions are common in cancer or in cancer-predisposing syndromes.1,2 This is exemplified by the observations that the normally repressed maternal IGF2 allele3,4 is frequently activated and the normally active maternal allele of the neighboring H19 gene5 is inactivated in a range of neoplasms, such as Wilms' tumors.6,7 Moreover, IGF2 is misregulated in patients with Beckwith–Wiedemann syndrome (BWS), a pathology associated with embryonic overgrowth of certain organs, skeletal abnormalities and a childhood predisposition to tumors.8 The interpretation of biallelically active IGF2 as a causative agent in diseases is supported by the finding that deliberate overexpression of IGF2 in mouse recapitulates the majority of the phenotypes associated with BWS.9

Current strategies to correct epigenetic lesions are severely hampered by the unspecific performance of drugs designed to modify epigenetic marks. For example, azacytidine restores DNA methylation-free domains in a genome-wide manner,10 and consequently has a range of unwanted effects. In order to examine opportunities for a more specific and efficient intervention, we have applied our rationale to target genes for activation and repression by synthetic zinc-finger transcription proteins (ZFPs) with novel binding specificities.11 Natural ZFPs employ tandem arrays of fingers that bind to a succession of neighboring subsites through interactions within the major groove of the DNA, wherein each finger makes precise contacts with a sequence of 3–4 bp. We, and others, have constructed libraries of such rationally designed and selected zinc-fingers that bind to a vast range of different subsites,12,13,14,15,16 and can use these for the assembly of synthetic ZFPs that recognize desired target sites within reporter genes.17 Furthermore, these approaches have been successfully applied to the regulation of endogenous genes both in the activation of genes, for example EPO1,18 VEGF-A,19 erbB-2 and erbB-3,13,20 and in the repression of gene expression, for example MDR1,21,22 erbB-2 and erbB-3,13,20 and PPARg.23

Here we show that an engineered ZFP can both positively and negatively regulate the expression of the IGF2–H19 genes. These effects depend upon both the functional domain employed and the direct binding of the ZFP to sites within the appropriate promoter. Importantly, we show that the transcriptionally inert alleles of both H19 and IGF2 are reactivated by the ZFP, documenting the ability to override epigenetically repressed states.


Activation of IGF2 using engineered zinc-finger transcription factors

The sequence of the entire IGF2 promoter region was scanned for 9-bp recognition sequences bound by a library of validated three-finger ZFP transcription factors.12 From the large set of molecules able to interact with this DNA region, five different synthetic IGF2-ZFPs were chosen with the following criteria: a binding site within ±500 bp of one or more of the four IGF2 promoters (Figure 1a), and an apparent dissociation constant of <100 pM, as determined by EMSA.19 The consensus-binding sites and apparent dissociation constants for these five proteins are given in the Materials and methods. To specify gene activation, these DNA binding moieties were fused to the VP16 activation domain. Figure 1e shows the protein sequence of one of these chimeric transcription factors in full. Transfection of human embryonic kidney cells (HEK293) with plasmid DNAs encoding these transcription factors increased the level of IGF2 mRNA from 1.8- to 50-fold, when compared to cells transfected with a plasmid containing no zinc-finger domain (Figure 1b).

Figure 1

Activation of IGF2 or H19 expression by ZFPs. (a) Schematic map of the IGF2 and H19 transcriptional units. Diamonds indicate the potential binding sites for ZFP809 found within ±500 bp of each start site of transcription. (b) IGF2 mRNA from transfected HEK293 cells was assayed by real-time PCR after expression of the various IGF2-ZFPs for 48 h. The IGF2 mRNA levels were quantified relative to an internal control of GAPDH mRNA and are shown as fold increase relative to the VP16 plasmid transfection (its value is set to 1.0). Each ZFP is numbered arbitrarily according to its library designation. (c) Activation of IGF2 expression by ZFP809 is domain dependent. The chart shows fold activation relative to the VP16 plasmid transfection (see panel b). (d) Activation of H19 expression by ZFPs. RNA isolated from HEK293 cells transfected with the indicated plasmids were measured for the steady-state H19 mRNA by real-time PCR with probes specific for H19 (Table 1) and normalized against the levels of GAPDH mRNA. The chart shows fold activation relative to controls as above. (e) The complete protein sequence of ZFP588-VP16 is shown indicating the domains relevant to in vivo function. Specifically, the sequences boxed and labeled finger 1, 2 and 3 contain the residues altered to confer alternative DNA binding specificities.

The highest level of activation was achieved by ZFP809, which also exhibits high affinity for its DNA recognition site in vitro (Kd 30 pM). To provide support for the direct function of this ZFP in the activation of IGF2, the DNA binding domain alone (no functional domain), and fusions of this DNA binding module to a second activation domain, p65 (derived from the RelA subunit of NF-κB), were tested as above (Figure 1c). While the chimera of the p65 activation domain with ZFP809 was able to activate IGF2 mRNA to levels indistinguishable from that of VP16, this activation was absolutely dependent on the presence of a functional domain as the DNA binding domain alone failed to regulate IGF2 mRNA levels (Figure 1c).

ZFP809 activates all four IGF2 promoters

Since the ZFP809 protein has a consensus-binding site within ±500 bp of the site of transcription initiation for each of the four IGF2 promoters, we performed RT-PCR experiments employing primer sets that differentiate between the four alternative promoter-specific IGF2 transcripts (see Methods and materials). Confirming previously published observations on IGF2 expression,24,25,26 the expression pattern for these promoters in HEK293 cells was characterized by P3 being the most significant transcript, whereas a somewhat weaker but substantial signal was observed for P4 (Figure 2, lanes 5 and 7). Promoters P1 and P2 are, to a first approximation, silent in this line (Figure 2, lanes 1 and 3). Following ZFP809 transfection, an intense elevation in the levels of the P3 and P4 promoter-specific transcripts was observed (Figure 2, lanes 6 and 8). Interestingly, the previously dormant P1 and P2 promoters are also activated to levels comparable to that of P4 in the absence of the ZFP (Figure 2, lanes 2 and 4). Thus, the increase in IGF2 mRNA realized upon transfection with ZFP809 results from the activation of all four promoters.

Figure 2

ZFP809-VP16 drives activation of all four IGF2 promoters. RNA isolated from HEK 293 cells transfected either with (ZFP) ZFP809-VP16 or non-transfected control cells (Ctrl) was analyzed by RT-PCR using primer sets specific for the different IGF2 transcripts (see Methods and materials, and Table 1). Equal amounts of RNA were used in each reaction (normalized by real-time PCR of GAPDH mRNA, not shown).

Targeted activation of H19, a putative tumor suppressor gene

Although the upstream promoter of H19 did not contain a binding site, a site for ZFP809 was found to be present downstream of the H19 transcription start site (Figure 1a). The transfection experiments described above (Figure 1c) were reanalyzed, therefore, using real-time RT-PCR with primers and probes directed against the H19 gene (see Table 1). Figure 1d shows that both the VP16 and p65 derivatives of ZFP809 drive a 40- to 120-fold increase in H19 expression levels, respectively, and that this upregulation was dependent upon the presence of a functional domain. The ZFP809-mediated activation of H19 is most simply explained by the direct action of the ZFP binding to a site proximal to its promoter. This observation (see also below) represents an interesting potential new strategy in cancer therapy, since the H19 gene, which is a putative tumor suppressor gene,27 is silenced during early stages of Wilms' tumorigenesis.28

Table 1 Primer and probe sequences employed

ZFP-driven repression of IGF2 and H19

Given that overexpression of IGF2 is associated with a number of disease states, it also would be of considerable therapeutic value to repress the expression of IGF2. To this end, the DNA recognition domain employed above was linked to the v-ErbA transcriptional repressor protein. v-ErbA is a viral form of the avian thyroid hormone receptor (TR), which is locked into its repressing state due to its inability to interact with ligand.29,30,31 The v-ErbA domain catalyzes the repression of transcription though associated histone deacetylation activity via its interactions with the HDAC3/NCoR co-repressor complex. Previous studies have used the KRAB repression domain to specify gene repression by ZFP-TFs,17,21 and a comparison of this domain with v-ErbA will be published in detail elsewhere (PDG and YJ, in preparation). Owing to the low basal levels of expression of IGF2 and H19 in HEK293 cells, we chose a carcinoma-derived cell line, U2OS, which expresses a significantly higher level of both genes, thus expanding the observable range for potential repression effects. Stable lines of U2OS were established, which express the ZFP transcription factor under the control of a tetracycline repressor (TetR) regulated CMV promoter. The addition of doxycycline (DOX) to these cells releases the CMV promoter from TetR repression, allowing regulated expression of the ZFP transcription factor.

Treating cells using a range of DOX concentrations from 0 to 2 ng/ml, over a 48 h period, induces increasing amounts of ZFP809-vErbA expression, demonstrating the proper and timely regulation of ZFP expression using the DOX inducible regulatory system (data not shown). Figures 3a and 3b shows that the steady-state levels of IGF2 and H19 RNA are strongly repressed under these conditions. At the maximal amount of DOX tested (2 ng/ml), H19 was repressed to 11% of its starting level in noninduced cells (Figure 3b), and IGF2 RNA levels were reduced to 20% of the starting value (Figure 3a). Indeed, the level of repression observed for both genes correlates well with the increasing levels of DOX administered. This repression showed dependence on the v-ErbA repression domain since minimal effects were observed upon IGF2 or H19 using a construct without functional domain (see Figures 1c and d). This result was observed for the vast majority of individual isolated clones, thus confirming the generality of this effect (data not shown). Taken together, these data show that ZFP809-vErbA is a potent, and domain dependent, transcriptional repressor of both the IGF2 and H19 promoters.

Figure 3

Repression of IGF2 and H19 by ZFP809-vErbA. Stably transformed U2OS cells containing copies of the ZFP809-vErbA gene under the control of a Tet-responsive full-length CMV promoter were grown for 48 h in the presence of DOX (concentration as indicated). RNA isolated from each condition was analyzed for (a) IGF2 or (b) H19 mRNA by real-time PCR (Taqman) and normalized against GAPDH mRNA as in Figure 1c. The noninduced sample is used as a reference and its value is arbitrarily set to 100%.

ZFP809 interacts directly with the IGF2 and H19 promoters

To demonstrate directly the binding of ZFP809 to the consensus-binding sites present in the IGF2 and H19 promoter regions, we performed chromatin immunoprecipitation (ChIP) experiments, employing the Flag epitope carried by the ZFP transcription factor. Using aliquots of the identical cells used in the repression assays above, the amount of DNA immunoprecipitated by the Flag antibody was quantified by real-time PCR using primers specific for each binding site (IGF2 promoters P1 and P3 have a prohibitively high GC content preventing direct analysis of binding at these two sites). Figure 4 shows that as the DOX concentration was increased, a selective enrichment for the IGF2 promoters P4, P2 (not shown), and H19 downstream promoter region was observed. The greatest extent of enrichment for each promoter was seen at the highest DOX concentration used (2 ng/ml), which also correlates with the most robust repression for both IGF2 and H19 (Figure 3). Two independent control fragments located at the GAPDH and VEGF-A loci that do not contain consensus-binding sites for ZFP809 demonstrated no significant enrichment upon ZFP induction (GAPDH results are shown in Figure 4). Furthermore, preimmune serum did not enrich for these promoter fragments (data not shown). These results confirm the in vivo binding of ZFP809 to the proposed binding sites, thus confirming the in vitro site selection data.12

Figure 4

ZFP809 acts directly at the IGF2 and H19 promoters in vivo. (a) ZFP809-vErbA protein binds to the IGF2 and H19 promoters in vivo. The stably transfected U2OS cell line containing copies of the ZFP809-vErbA gene described in Figure 3, was induced for 48 h with the indicated DOX concentrations, and the in vivo binding of the ZFP to the endogenous IGF2 and H19 promoters analyzed by chromatin immunoprecipitation using an antibody against the Flag epitope of the ZFP protein (see Methods and materials). Three primer sets were used to monitor the binding of the ZFP to the IGF2 promoter 4, H19 promoter and, as a control, the GAPDH exon 7 coding sequence. Results are expressed as the fold increase relative to the same cells in the absence of DOX, and were corrected for differences in the amount of starting material (input). No enrichment was observed with preimmune serum (data not shown). (b) ZFP809-TR directly activates the IGF2 and H19 promoters in vivo. The stably transfected U2OS cell line containing copies of the ZFP809-TR gene was induced for 24 h by the addition of 1 ng/ml DOX, the media was supplemented with 100 μg/ml of the protein synthesis inhibitor cycloheximide for 2 h (as indicated), and incubated for a further 24 h following the further addition (100 μg/ml) or absence of activating ligand (T3) as indicated. IGF2 and H19 mRNA levels were quantified relative to an internal control of GAPDH mRNA by real-time PCR (Taqman) as described in Figure 1, and are shown as fold change relative to the identically treated cell line in the absence of doxycycline (ie no ZFP-TF).

To further support the direct action of our ZFP transcription factors on the IGF2 and H19 genes, we experimentally determined whether these targets were direct or secondary effects of ZFP regulation. To do this we employed ZFP809 fused to the chicken thyroid receptor (alpha 1) ligand-binding domain (aa 628–1512), the closest endogenous relative of v-ErbA. This domain places the functional activity of the ZFP under small molecule control such that ZFP809-TR in the absence of ligand (−T3) is a transcriptional repressor (analogous to v-ErbA), whereas in the presence of hormone (+T3) this chimera is a transcriptional activator (see Figure 4b). This activity is dependent upon the ZFP DNA binding domain since a GFP-TR fusion is unable to regulate the expression of IGF2 irrespective of the presence or absence of hormone in transient transfection assays (data not shown). To differentiate the direct from indirect effects of the ZFP fusion, we employed the protein synthesis inhibitor cycloheximide (CHX). In the absence of new protein synthesis, the mRNA expression of only those genes that are direct targets of the ZFP chimera will continue to be regulated. Stable lines of U2OS were established which express this ZFP-TR transcription factor under the control of a tetracycline repressor (TetR) regulated CMV promoter. The addition of DOX to these cells releases the CMV promoter from TetR repression, allowing regulated expression of the ZFP transcription factor. Following the provision of fresh media, DOX was added to induce ZFP expression, 24 h after DOX addition protein synthesis was inhibited by the addition of CHX to the media, and 2 h later ligand (+T3) was added to the media or not (−T3) and the cells incubated for a further 24 h. The concentration of cycloheximide used in these assays is identical to published work using CHX with this cell type,32 and is 10-fold higher than that necessary to prevent de novo protein synthesis in U2OS cells.33 The results of this experiment (Figure 4b) show that even in the absence of new protein synthesis, we continue to observe an upregulation of both the IGF2 and H19 genes. Neither CHX alone nor T3 alone are able to invoke an activation of either target gene (Figure 4b and data not shown). We note that the presence of CHX leads to an increase in the fold activation of H19 relative to the control. The increased fold change is a result of the reduction in basal H19 RNA levels caused by CHX treatment, and does not reflect a higher absolute induced level of H19 per se. Taken together, the ChIP binding data and the upregulation of both IGF2 and H19 in the absence of protein synthesis confirm that these genes are true direct targets of the ZFP-chimeras employed.

Reactivation of the silent IGF2 and H19 alleles

Both in terms of providing insights into the process of imprinting and as a test of the engineered transcription factor technology, we wished to determine if the ZFP809 protein was able to overcome the repressed states of IGF2 and H19. To this end, we exploited a restriction site polymorphism located in the exon V of the H19 gene in cultured HEK293 cells, which maintain monoallelic expression of H19 and IGF2.34 As a control we confirmed that only one allele was expressed in nontransfected HEK293 cells (Figure 5a, lanes 3 and 4). Next, we repeated the experiment using RNA preparations from HEK293 cells transfected with ZFP809 linked to VP16. The results shown in Figure 5a (lanes 1 and 2) demonstrate that the ZFP strongly increases the expression of the active allele. Importantly, however, this engineered transcription factor is also able to reactivate the expression of the silent allele of H19 (Figure 5a, upper band), showing that an engineered ZFP overrides the imprinted transcriptional status of the H19 promoter by reactivating the nonexpressed paternal allele.

Figure 5

ZFP-driven reactivation of the transcriptionally silent IGF2 and H19 alleles. (a) RNA isolated from HEK293 cells transfected with ZFP809-VP16 (indicated) or a VP16 control plasmid (VP16 plasmid) were analyzed by RT-PCR using the H19 specific RsaI-F and RsaI-R primers (see Methods and materials, Table 1). The PCR products were then digested overnight with RsaI, resolved on an agarose gel, which was blotted and probed with radiolabelled RsaI-Forward primer. RNA loading was normalized by real-time PCR analysis of 18S (relative units given below lanes). (b) ZFP-driven reactivation of the silent IGF2 and H19 alleles in mouse–human cell hybrids containing a single copy of human chromosome 11 of known parental origin. RNA isolated from P11PA9 (paternal) or M11PA9 (maternal) mouse–human cell hybrids transfected with ZFP809-VP16 were analyzed by RNAse protection assays for the expression of both IGF2 and H19 and compared to untransfected and GFP-VP16 control plasmids. Lanes 2–6 are for the paternally derived chromosome and lanes 7–11 for the maternal, while lane 1 is the result for no input RNA. The results for duplicate transfection experiments are indicated (exp. 1 or exp. 2).

To provide additional support for the activation of the nonexpressed alleles, we employed murine cell hybrids that contain a single human chromosome of known parental origin.35 These cell hybrids retain the monoallelic expression patterns for both IGF2 and H19 (Figure 5b, lanes 2 and 7), and thus provide a powerful tool for the study of each chromosome in isolation.36 We transfected these cells with ZFP809-VP16 or GFP-VP16 plasmids, and analyzed the expression of both genes by RNAse protection assays. The results of these experiments confirm our previous analysis of H19 expression, and we observe the reactivation of the silent allele of H19 in the cell line bearing only the paternal copy of chromosome 11 (Figure 5b, lanes 3 and 4). Moreover, in the cell line bearing the maternal chromosome 11, and thus the silent allele of IGF2, transfection of plasmid encoding the ZFP809-VP16 chimera resulted in the dramatic reactivation of IGF2 (Figure 5b, lanes 8 and 9). These data demonstrate that designed transcription factors can override the epigenetically silenced states of two imprinted genes.


We have shown here that specific VP16 or p65 activation domain-containing ZFP chimeras dramatically activate both IGF2 and H19. Exchanging the functional domain for a repression domain (v-ErbA) reversed this phenotype, and resulted in strong downregulation of both genes. The domain-dependent, reversible nature of the transcriptional response, combined with the cycloheximide and ChIP experiments that demonstrate the interaction of the ZFP with target DNA in the promoters of both genes, strongly supports interpretation of these results as direct effects. Central to this report is our observation that ZFP809 is able to reactivate the transcriptionally inert alleles of both IGF2 and H19. An engineered zinc-finger transcription factor can hence override the imprinted status of both the IGF2 and H19 genes.

ZFP-driven control of IGF2 refines models for genomic imprinting

Monoallelic expression of imprinted genes is the consequence of the proper orchestration of a complex, and only partially understood process. Primary to this phenomenon is the imprinting control region located in the 5′-flank of H19, which prevents the enhancers located 3′ to the H19 gene from interacting with the IGF2 promoters some 80 kb away on the maternal chromosome, while silencing the H19 gene on the paternal chromosome.37 Mechanistically, these processes remain poorly understood although the repression of the maternally derived IGF2 may involve the chromatin insulator protein CTCF,38 which interacts with the H19 ICR in an allele-specific39 and methylation-sensitive manner.39,40,41

The fact that a ZFP transcription factor can bind to the IGF2 promoter and activate expression of the transcriptionally silent allele of IGF2 rules out models for insulator function in which the IGF2 promoter region is held within a repressive chromatin architecture, or is compartmentalized into a region or zone of the nucleus recalcitrant to the process of transcription. On the contrary, all four IGF2 promoters (spanning some 20 kb of chromosome 11) appear highly amenable to transcriptional upregulation. The loss of IGF2 imprinting by the simple provision of local ‘enhancer’ function, in the form of a transcriptional activation domain, suggests that an activation signal was the sole limiting factor preventing expression. Our results therefore support models in which the CTCF insulator activity provides the primary determinant for the activity of IGF2 by continuously regulating the interaction of the enhancers with the IGF2 promoter.

Stringency of the repressed state of the paternal H19 allele revisited

There is little reason to assume that the repressed states of the maternal IGF2 and paternal H19 alleles are identical or even similar. While the maintenance of a repressed maternal IGF2 allele depends on the continuous presence of the H19 ICR, the conditional deletion of the paternal H19 ICR allele does not lead to H19 activation.42 Taken together, the data indicate that the repressed status of the paternal H19 allele depends on promoter methylation that originates from the methylated H19 ICR during early development. The actual silencing of the paternal H19 allele has been proposed to depend on the methylation-dependent recruitment of histone deacetylase43 via MeCP2.44 However, such observations fail to acknowledge the finding that the H19 paternal allele is actively transcribed, at least in mouse cells, although the stability of its transcript is greatly impaired resulting in the apparent expression of only the maternal allele.45 Moreover, our ChIP data show that the proximal promoter of H19 is permissive to the binding of a transcription factor, and that this binding can result in the upregulation of the silent allele. We note, however, that the ZFP used in these studies recognizes a binding site that does not contain a CpG dinucleotide and therefore is unlikely to be directly affected by the methylation status of the region – a feature that is exploited by CTCF – and potentially other natural factors necessary for the maintenance of the imprinted state. That said, the argument that a generally repressive chromatin structure prevents expression of the paternal H19 is shown here to be inaccurate.

Applications and therapeutic implications

The data described here demonstrate the ability of an engineered transcription factor to enforce control over the imprinted transcriptional status and highlight the potential of synthetic transcription factors in the treatment of epigenetic lesions. Our demonstration that the interaction between these designed ZFPs and their target sites ignores different types of epigenetically repressed states suggests that any gene can be activated or repressed at will. Apart from correcting imprinting states in diseases exhibiting loss of imprinting, therapeutic targets for synthetic ZFPs include activation of silenced, nonimprinted tumor suppressor genes and suppression of angiogenesis. In addition, the synthetic transcription factors might be exploited to modulate stem cell differentiation to generate insulin-producing cells, for example (see Wu et al46; Kim and Pabo47; Reik48 and references therein). Key to the success of artificial transcription factors as therapeutic interventions is the specificity of ZFP action. The proteins described herein bind 9-bp recognition sites, which given a random genome, would be predicted to bind to many thousands of locations. However, the restriction on promiscuous binding enforced by the chromatin architecture of the genome,18,19 the requirement for correct location of binding sites within regulatory regions such as promoters or enhancers, and the short functional range of the targeted effecter domains may result in sufficient specificity of function. That said, the modular nature of the zinc-finger binding domain makes possible the construction of multifinger transcription factors with four, or more, such domains20,21,49 providing increased specificity such that a single gene is regulated genome-wide. As our knowledge of the molecular basis of disease improves, and the key regulators involved are identified, the ability to apply such transcription-based therapies will be of increasing clinical utility.

Note added to proof

Recent work has linked in vitro fertilization with Beckwith-Wiedemann Syndrome – a human overgrowth syndrome associated with epigenetic alterations of the imprinted IGF2, H19 and LIT1genes. This highlights the importance of treatments for epigenetic disease. [Association of In Vitro Fertilization with Beckwith-Wiedemann Syndrome and Epigenetic Alterations of LIT1 and H19 Michael R. DeBaun, Emily L. Niemitz and Andrew P. Fernberg Am J. Hum. Genet., 72. published online 11/19/2002.] Moreover, the data described herein (using designed zinc-finger proteins) provide the molecular foundation for a transcription factor-mediated corrective gene therapy.


We thank Andrew Jamieson, Edward Rebar and Qiang Liu for the design and selections of the ZFPs in the library, Monica Miller and Matthew Mendel for assembly and initial testing of the ZFP constructs, Xiaohong Zhong for technical help in the generation of the stable cell line, and Fyodor Urnov for comments on the manuscript. We are also grateful to Edward Lanphier, Casey Case, Carl Pabo and Elizabeth Wolffe for encouragement and support. This work was supported in part by grants from the Swedish Science Research Council (VR), the Swedish Cancer Research Foundation (CF), the Swedish Pediatric Cancer Foundation (BCF) and the Lundberg Foundation all to RO, and NIH grant CA54358 to APF. We dedicate this work to the memory of our friend, colleague and mentor Alan P Wolffe.

Experimental procedures

Synthesis, purification and gel shift analysis of zinc-finger proteins

Genes encoding our IGF2 targeted ZFPs were assembled, cloned and purified as previously described.18 Briefly, oligonucleotides encoding α-helix and β-sheet regions of each three-finger protein were assembled using PCR, and each resultant ZFP gene was cloned into the pMal-c2 plasmid (New England Biolabs Inc.) as a fusion with DNA encoding maltose-binding protein. Maltose-binding protein–ZFP fusions were then expressed and affinity-purified using an amylose resin (New England Biolabs Inc.).

Binding studies were performed essentially as described.19 The ZFPs employed in this study have the following dissociation constants and consensus-binding sites: ZFP240 (G(G/a)GGAGGC(A/t): Kd100 pM), ZFP588 (A(G/a)GGTGGAC: Kd30p M), ZFP727 (G(A/g)TG(G/a)GGAG: Kd25 pM), ZFP735 (G(C/t)GGTGGCT: Kd50 pM) and ZFP809 (G(G/c)GGG(T/a)G(A/g)C: Kd30 pM). Consensus sequences (either empirically determined or predicted from selection experiments) indicate the specific subset of tolerated single base changes (lowercase) which alone are predicted or known to result in a mild (

ZFP expression constructs used for cell culture studies

ZFPs with distinct binding properties were assembled and cloned into the pcDNA3 mammalian expression vector (Invitrogen) as described previously.18 A cytomegalovirus promoter drove the expression of all the ZFPs in mammalian cells. All ZFP constructs contained an N-terminal nuclear localization signal (Pro-Lys-Lys-Lys-Arg-Lys-Val) from SV40 large T antigen, a ZFP DNA binding domain, a functional domain where indicated, and a FLAG peptide (Asp-Tyr-Lys-Asp-Asp-Asp-Asp-Lys). ZFP-VP16 fusions contained the herpes simplex virus VP16 activation domain from amino acids 413–490.18,50. ZFP-p65 fusions contained the human NF-κB transcription factor p65 subunit (amino acids 288–551) as the activation domain.51 ZFP-vErbA fusions contained the complete C-terminal portion (aa 270–555) of the v-ErbA transcriptional repressor.29,30 ZFP-TR fusions contained the complete C-terminal portion (aa 114–408) of the chicken TR alpha 1 (c-ErbA).52

Assay for activity of ZFP fusions for endogenous gene activity in human cells by transient transfection

HEK293 cells were plated in 24-well plates at a density of 160 000 cells/well and were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum in a 5% CO2 incubator at 37°C. Plasmids encoding ZFP-VP16, ZFP-p65 or ZFP-vErbA fusions were transfected into the cells 1 day later, via LipofectAMINE 2000 reagent (Gibco-BRL, MD, USA) according to the manufacturer's recommendations, using 1.5 μl of LipofectAMINE reagent and 0.3 μg of ZFP plasmid DNA per well. The medium was removed and replaced with fresh medium 16 h after transfection. The culture medium and the cells were harvested and assayed for the expression of the gene of interest by RT-PCR 48 h after transfection. Transfection efficiency was controlled in HEK293 cells by transfecting a GFP plasmid expression plasmid in the same experiment. After 48 h, the percentage of GFP-positive cells was 80–90% for all experiments analyzed.

Quantitative real-time RT-PCR analysis of IGF2/H19 mRNA levels

Cells were lysed, and total RNA was prepared using the RNeasy total RNA isolation kit with in-column DNase treatment (QIAGEN Inc.). RNA (25 ng) was used in real-time quantitative RT-PCR analysis using Taqman chemistry in a 96-well format on an ABI 7700 SDS machine (PerkinElmer Life Sciences) as described previously.18 Briefly, reverse transcription was performed at 48°C for 30 min using MultiScribe reverse transcriptase (PerkinElmer Life Sciences). Following a 10-min denaturation at 95°C, PCR amplification using AmpliGold DNA polymerase was conducted for 40 cycles at 95°C for 15 s and at 60°C for 1 min. The results were analyzed using SDS Version 1.6.3 software. The primers and probes used for Taqman analysis are listed in Table 1.

Standard quantitative RT-PCR

The Thermoscript™ RT-PCR system (Invitrogen, CA, USA) was used according to the manufacturer's recommendations.The reverse transcription was performed at 48°C for 30 min. After denaturing at 95°C for 10 min, PCR amplification reactions were conducted for 35 cycles at 95°C for 15 s and at 60°C for 1 min. The entire sample was then loaded onto a 3% agarose gel, before electrophoresis and transfer to a Nytran membrane (Schleicher and Schuel, Keene, NH). After end labeling, an oligonucleotide probe was used for detection of the four different transcripts (IGF2-P: IndexTermTGAGAAGCACCAGCATCGACTTCCCCATT). Hybridization was carried out in Rapid Hyb buffer (Amersham, MA, USA). Hybridization and subsequent washings were performed at 65°C. The results were subsequently visualized by a PhosporImager.

For H19 allele specific expression studies, the reverse transcription was performed using the same conditions as above except that the PCR amplification was conducted for 30 cycles at 95°C for 30 s, 30°C for 30 s and at 72°C for 2 min. The primers used are shown in Table 1. The recovered DNA was digested to completion overnight at 37°C with the RsaI restriction endonuclease and then concentrated by ethanol precipitation. The samples were subsequently loaded on a 3% agarose gel, blotted on to Nytran membrances, and hybridiz as for IGF2 above, except that the RsaI-F oligonucleotide was used as the probe.

Generation of stable inducible cell lines

The T-REx™-U2OS cell line was purchased from Invitrogen and cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum in a 5% CO2 incubator at 37°C. To generate stable Tet-inducible cell lines expressing the ZFP809-vErbA transcription factor, the coding region from the pZFP809-vErbA vector was subcloned into pcDNA4/TO (Invitrogen) using the AflII and XhoI restriction sites. The resulting pTO-ZFP809-vErbA construct was transfected into the T-REx-U2OS™ cell line using LipofectAMINE 2000 (Life Technologies, Inc.). After 2 weeks of selection in medium containing 400 μg/ml Zeocin™ (Invitrogen), individual clones were isolated and analyzed for doxycycline-dependent expression of ZFP809-vErbA expression and corresponding repression of the endogenous gene target.

Chromatin immunoprecipitation was performed using a ChIP assay kit according to the manufacturer's instructions (Upstate Biotechnology, NY, USA). Briefly, approximately five million cells were crosslinked with 1% formaldehyde for 10 min, washed with PBS, and resuspended in lysis buffer. The cell lysate was sonicated on ice, resulting in an average DNA fragment length of approximately 500 bp. After removing cell debris by centrifugation, immunoprecipitation was performed in ChIP dilution buffer overnight with anti-Flag M2 affinity gel (Sigma, Germany), diluted at 1/20. The antibody–agarose complex was centrifuged and washed, and the immunoprecipitated fraction eluted. The crosslinking was reversed by incubation at 65°C for 4 h in the presence of 200 mM NaCl. The DNA was recovered by phenol/chloroform extraction, and the abundance of specific sequences quantitated using real-time PCR as described above.

Maintenance, transfection and RNAse protection assays in mouse–human hybrid cell lines

The M11PA9 maternal and P11PA9 paternal cell lines were maintained in Dulbecco MEM (Gibco BRL), supplemented with 10% FBS, penicillin/streptomycin (Gibco) and Blasticidin S (CAYLA). Transfections were carried out using FuGene 6 transfection reagent (Roche) with equimolar amounts of the plasmid DNAs (5 μg of ZFP plasmid DNA) and 0.5 μg of a reference plasmid containing a β-galactosidase reporter gene under the control of the SV40 promoter-enhancer (pSVβGal), as has been described.53

For the RNAse protection assay of mouse–human hybrid cell lines – RNA was extracted as described previously and subjected to RNAse protection assay using the Ambion RPA III kit as per the manufacturer's instructions. Dde I-linearized human H19 cDNA cloned into pBluescript was used to generate an H19-specific antisense riboprobe (253 bases; T7 polymerase). A 558 bp HinfI-PstI human IGF2 cDNA insert, cloned in pGEM-3 and encompassing exon 7 and the 5′-region of exon 9, was digested with XhoI and used as the template for the generation of an IGF2 antisense riboprobe (145 bases; SP6 polymerase). Relative gene expression was visualized using a Fuji FLA 3000 Phosphorimager.


  1. 1

    Feinberg A, Vogelstein B . Hypomethylation distinguishes genes of some human cancers from their normal counterparts. Nature 1983; 301: 89–92.

    CAS  Article  Google Scholar 

  2. 2

    Robertson K . DNA methylation, methyltransferases, and cancer. Oncogene 2001; 20, 3139–3155.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  3. 3

    Ohlsson R et al. IGF2 is parentally imprinted during human embryogenesis and in the Beckwith–Weidemann syndrome. Nature Genet 1993; 4: 94–97.

    CAS  PubMed  Article  Google Scholar 

  4. 4

    Giannoukakis N, Deal C, Paquette J, Goodyer C, Polychronakos C . Parental genomic imprinting of the human IGF2 gene. Nature Genet 1993; 4: 98–101.

    CAS  PubMed  Article  Google Scholar 

  5. 5

    Zhang Y, Tycko B . Monoallelic expression of the human H19 gene. Nature Genet 1992; 1: 40–44.

    CAS  PubMed  Article  Google Scholar 

  6. 6

    Rainier S et al. Relaxation of imprinted genes in human cancer. Nature 1993; 362: 747–749.

    CAS  Article  Google Scholar 

  7. 7

    Ogawa O et al. Relaxation of insulin-like growth factor II gene imprinting implicated in Wilms' tumour. Nature 1993; 362: 749–751.

    CAS  Article  Google Scholar 

  8. 8

    Weksberg R, Shen DR, Fei YL, Song QL, Squire J . Disruption of insulin-like growth factor 2 imprinting in Beckwith–Wiedemann syndrome. Nature Genet 1993; 5: 143–150.

    CAS  PubMed  Article  Google Scholar 

  9. 9

    Sun FL, Dean WL, Kelsey G, Allen ND, Reik W . Transactivation of Igf2 in a mouse model of Beckwith–Wiedemann syndrome. Nature 1997; 389: 809–815.

    CAS  PubMed  Article  Google Scholar 

  10. 10

    Eversole-Cire P et al. Activation of an imprinted Igf 2 gene in mouse somatic cell cultures. Mol Cell Biol 1993; 13: 4928–4938.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  11. 11

    Pabo CO, Peisach E, Grant RA . Design and selection of novel Cys2His2 zinc finger proteins. Annu Rev Biochem 2001; 70: 313–340.

    CAS  PubMed  Article  Google Scholar 

  12. 12

    Liu Q, Xia Z, Case CC . Validated zinc finger protein designs for all 16 GNN DNA triplet targets. J Biol Chem 2002; 277: 3850–3856.

    CAS  PubMed  Article  Google Scholar 

  13. 13

    Dreier B, Beerli RR, Segal DJ, Flippin JD, Barbas CF 3rd . Development of zinc finger domains for recognition of the 5′-ANN-3′ family of DNA sequences and their use in the construction of artificial transcription factors. J Biol Chem 2001; 276: 29466–29478.

    CAS  PubMed  Article  Google Scholar 

  14. 14

    Segal DJ, Dreier B, Beerli RR, Barbas CF 3rd . Toward controlling gene expression at will: selection and design of zinc finger domains recognizing each of the 5′-GNN-3′ DNA target sequences. Proc Natl Acad Sci USA 1999; 96: 2758–2763.

    CAS  PubMed  Article  Google Scholar 

  15. 15

    Rebar EJ, Greisman HA, Pabo CO . Phage display methods for selecting zinc finger proteins with novel DNA-binding specificities. Methods Enzymol 1996; 267: 129–149.

    CAS  PubMed  Article  Google Scholar 

  16. 16

    Rebar EJ, Pabo CO . Zinc finger phage: affinity selection of fingers with new DNA-binding specificities. Science 1994; 263: 671–673.

    CAS  Article  Google Scholar 

  17. 17

    Beerli RR, Segal DJ, Dreier B, Barbas CF 3rd . Toward controlling gene expression at will: specific regulation of the erbB-2/HER-2 promoter by using polydactyl zinc finger proteins constructed from modular building blocks. Proc Natl Acad Sci USA 1998; 95: 14628–14633.

    CAS  Article  Google Scholar 

  18. 18

    Zhang L et al. Synthetic zinc finger transcription factor action at an endogenous chromosomal site. Activation of the human erythropoietin gene. J Biol Chem 2000; 275: 33850–33860.

    CAS  PubMed  Article  Google Scholar 

  19. 19

    Liu PQ et al. Regulation of an endogenous locus using a panel of designed zinc finger proteins targeted to accessible chromatin regions. Activation of vascular endothelial growth factor A. J Biol Chem 2001; 276: 11323–11334.

    CAS  PubMed  Article  Google Scholar 

  20. 20

    Beerli RR, Dreier B, Barbas CF 3rd . Positive and negative regulation of endogenous genes by designed transcription factors. Proc Natl Acad Sci USA 2000; 97: 1495–1500.

    CAS  PubMed  Article  Google Scholar 

  21. 21

    Bartsevich VV, Juliano RL . Regulation of the MDR1 gene by transcriptional repressors selected using peptide combinatorial libraries. Mol Pharmacol 2000; 58: 1–10.

    CAS  PubMed  Article  Google Scholar 

  22. 22

    Xu D, Ye D, Fisher M, Juliano RL . Selective inhibition of P-glycoprotein expression in multidrug-resistant tumor cells by a designed transcriptional regulator. J Pharmacol Exp Ther 2002; 302: 963–971.

    CAS  PubMed  Article  Google Scholar 

  23. 23

    Ren D, Collingwood TN, Rebar EJ, Wolffe AP, Camp HS . PPAR gamma knockdown by engineered transcription factors: exogenous PPARgamma2 but not PPARgamma1 reactivates adipogenesis. Genes Dev 2002; 16: 27–32.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  24. 24

    Ekström TJ, Cui H, Li X, Ohlsson R . Promoter-specific IGF2 imprinting status and its plasticity during human liver development. Development 1995; 121: 309–316.

    PubMed  Google Scholar 

  25. 25

    van Dijk MA, van Schaik FM, Bootsma HJ, Holthuizen JS . Initial charaterization of the four promoters of the human IGF II gene. Mol Cell Endocrinol 1991; 81: 81–94.

    CAS  PubMed  Article  Google Scholar 

  26. 26

    Vu TH, Hoffman AR . Promoter-specific imprinting of the human insulin-like growth factor-II gene. Nature 1994; 371: 714–717.

    CAS  PubMed  Article  Google Scholar 

  27. 27

    Hao Y, Crenshaw T, Moulton T, Newcomb E, Tycko B . Tumour-suppressor activity of H19 RNA. Nature 1993; 365: 764–767.

    CAS  Article  Google Scholar 

  28. 28

    Cui H et al. Inactivation of H19, an imprinted and putative tumor repressor gene, is a preneoplastic event during Wilms' tumorigenesis. Cancer Res 1997; 57: 4469–4473.

    CAS  PubMed  Google Scholar 

  29. 29

    Sap J, Munoz A, Schmitt J, Stunnenberg H, Vennstrom B . Repression of transcription mediated at a thyroid hormone response element by the v-erb-A oncogene product. Nature 1989; 340: 242–244.

    CAS  PubMed  Article  Google Scholar 

  30. 30

    Damm K, Thompson CC, Evans RM . Protein encoded by v-erbA functions as a thyroid-hormone receptor antagonist. Nature 1989; 339: 593–597.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  31. 31

    Zenke M, Munoz A, Sap J, Vennstrom B, Beug H . v-erbA oncogene activation entails the loss of hormone-dependent regulator activity of c-erbA. Cell 1990; 61: 1035–1049.

    CAS  PubMed  Article  Google Scholar 

  32. 32

    Yao F, Schaffer PA . An activity specified by the osteosarcoma line U2OS can substitute functionally for ICP0, a major regulatory protein of herpes simplex virus type 1. J Virol 1995; 69: 6249–6258.

    CAS  PubMed  PubMed Central  Google Scholar 

  33. 33

    Koziczak M, Muller H, Helin K, Nagamine Y . E2F1-mediated transcriptional inhibition of the plasminogen activator inhibitor type 1 gene. Eur J Biochem 2001; 268: 4969–4978.

    CAS  PubMed  Article  Google Scholar 

  34. 34

    Kiess W, Paquette J, Koepf G, Wolf E, Deal C . Proinsulin-like growth factor-II overexpression does not alter monoallelic H19 gene expression in transfected human embryonic kidney fibroblasts. Biochem Biophys Res Commun 1999; 255: 226–230.

    CAS  PubMed  Article  Google Scholar 

  35. 35

    Kugoh H et al. Mouse A9 cells containing single human chromosomes for analysis of genomic imprinting. DNA Res 1999; 6: 165–172.

    CAS  Article  Google Scholar 

  36. 36

    Inoue J et al. Construction of 700 human/mouse A9 monochromosomal hybrids and analysis of imprinted genes on human chromosome 6. J Hum Genet 2001; 46: 137–145.

    CAS  PubMed  Article  Google Scholar 

  37. 37

    Thorvaldsen JL, Duran KL, Bartolomei MS . Deletion of the H19 differentially methylated domain results in loss of imprinted expression of H19 and Igf2. Genes Dev 1998; 12: 3693–3702.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  38. 38

    Ohlsson R, Renkawitz R, Lobanenkov V . CTCF is a uniquely versatile transcription regulator linked to epigenetics and disease. Trends Genet 2001; 17: 520–527.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  39. 39

    Kanduri C et al. Functional interaction of CTCF with the insulator upstream of the H19 gene is parent of origin-specific and methylation-sensitive. Curr Biol 2000; 10: 853–856.

    CAS  PubMed  Article  Google Scholar 

  40. 40

    Bell AC, Felsenfeld G . Methylation of a CTCF-dependent boundary controls imprinted expression of the Igf2 gene. Nature 2000; 405: 482–485.

    CAS  Article  Google Scholar 

  41. 41

    Hark AT et al. CTCF mediates methylation-sensitive enhancer-blocking activity at the H19/Igf2 locus. Nature 2000; 405: 486–489.

    CAS  Article  Google Scholar 

  42. 42

    Kaffer CR et al. A transcriptional insulator at the imprinted H19/Igf2 locus. Genes Dev. 2000; 14: 1908–1919.

    CAS  PubMed  PubMed Central  Google Scholar 

  43. 43

    Svensson K et al. The paternal allele of the H19 gene is progressively silenced during early mouse development: the acetylation status of histones may be involved in the generation of variegated expression patterns. Development 1998; 125: 61–69.

    CAS  PubMed  Google Scholar 

  44. 44

    Drewell RA, Goddard CJ, Thomas JO, Surani MA . Methylation-dependent silencing at the H19 imprinting control region by MeCP2. Nucleic Acids Res 2002; 30: 1139–1144.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  45. 45

    Jouvenot Y, Poirier F, Jami J, Paldi A . Biallelic transcription of Igf2 and H19 in individual cells suggests a post-transcriptional contribution to genomic imprinting. Curr Biol 1999; 9: 1199–1202.

    CAS  PubMed  Article  Google Scholar 

  46. 46

    Wu H, Yang WP, Barbas CF 3rd. Building zinc fingers by selection: toward a therapeutic application. Proc Natl Acad Sci USA 1995; 92: 344–348.

    CAS  Article  Google Scholar 

  47. 47

    Kim JS, Pabo CO . Transcriptional repression by zinc finger peptides. Exploring the potential for applications in gene therapy. J Biol Chem 1997; 272: 29795–29800.

    CAS  Article  Google Scholar 

  48. 48

    Reik A, Gregory PD, Urnov FD . Biotechnologies and therapeutics: chromatin as a target. Curr Opin Genet Dev 2002; 12: 233–242.

    CAS  PubMed  Article  Google Scholar 

  49. 49

    Kim JS, Pabo CO . Getting a handhold on DNA: design of poly-zinc finger proteins with femtomolar dissociation constants. Proc Natl Acad Sci USA 1998; 95: 2812–2817.

    CAS  PubMed  Article  Google Scholar 

  50. 50

    Sadowski I, Ma J, Triezenberg S, Ptashne M . GAL4-VP16 is an unusually potent transcriptional activator. Nature 1988; 335: 563–564.

    CAS  Article  Google Scholar 

  51. 51

    Ruben SM et al. Isolation of a rel-related human cDNA that potentially encodes the 65-kD subunit of NF-kappa B. Science 1991; 251: 1490–1493.

    CAS  PubMed  Article  Google Scholar 

  52. 52

    Sap J et al. The c-erb-A protein is a high-affinity receptor for thyroid hormone. Nature 1986; 324: 635–640.

    CAS  Article  Google Scholar 

  53. 53

    Franklin G et al. The human PDGF-B gene is regulated by cell type-specific activator and suppressor elements in the first intron. EMBO J 1991; 10: 1365–1373.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to P D Gregory.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Jouvenot, Y., Ginjala, V., Zhang, L. et al. Targeted regulation of imprinted genes by synthetic zinc-finger transcription factors. Gene Ther 10, 513–522 (2003).

Download citation


  • zinc-finger proteins
  • gene regulation
  • imprinting
  • transcriptional repression

Further reading