ZFN Technology

Biowire Volume 10 Article 1

Have your genomic work cut out for you

The genomes of several organisms, including humans, have been sequenced, yet we have a limited understanding of the functional implications of the identified genes.

Extensive analysis is required to understand the function of each gene and put that into the systemic context. This has led to an increased focus on functional genomics. Traditional approaches to understand gene function in mammalian systems have been limited to the use of RNAi or genetically modified mice, due to the lack of requisite tools for targeted genome editing in other systems. The discovery of the modularity of zinc finger proteins and subsequent development of Zinc Finger Nuclease (ZFN) technology has trampled these technological barriers, giving scientists the ability to manipulate the genome of any organism in a highly targeted manner. The applications of ZFNs are broad and far-reaching, enabling scientists to answer technologically challenging questions with unparalleled ease.

What are Zinc Finger Nucleases?

Zinc Finger Nucleases are a class of engineered proteins that create a highly targeted double-strand break (DSB) within the genome and enable the manipulation of the genome with unprecedented ease and precision. ZFNs consist of two domains, a zinc finger DNA binding domain and a catalytic nuclease domain (Figure 1). The zinc finger domains used here are derived from the Cys(2)-His(2) zinc finger proteins that each recognize 3 bases of DNA. These zinc finger proteins can be engineered to recognize most 3 base pair sequences, allowing targeting to almost any sequence in the genome. The modularity of each zinc finger permits us to link together 4-6 zinc finger proteins creating 12-18 base pair specificity within the genome.

Figure 1 Each Zinc Finger Nuclease (ZFN) consists of two functional domains: A DNA-binding domain comprised of a chain of zinc finger modules, each recognizing a unique triplet (3 bp) sequence of DNA. Four to six zinc finger modules are stitched together to form a Zinc Finger Protein (ZFP), with specificity of ≥12 bp. A DNA-cleaving domain comprised of the nuclease domain of FokI is attached to the ZFPs. When the DNA-binding and DNA-cleaving domains are fused together, a highly specific pair of ‘genomic scissors’ is created that binds with 24-36 bp specificity of the ZFPs and cleaves the DNA.

Fewer fingers can also be used but specificity is reduced. To complete the engineered ZFN, we add a modular nuclease domain from FokI, a Type IIs restriction endonuclease (which has a separate DNAbinding domain and nuclease domain), to the zinc finger domain. The nuclease domain from FokI functions as a dimer, requiring that a second ZFN binds adjacent to the first one. The requirement of a pair of zinc finger nucleases to create a double-strand break gives the site 24-36 base pair specificity. The target sites of the two ZFNs must be separated by 5-7 base pairs to allow formation of the catalytically active FokI dimer (Figure 1).

How do Zinc Finger Nucleases work?

Zinc Finger Nucleases can be designed to target any gene in any genome. These can then be delivered to the cell as DNA or RNA. The ZFN proteins are expressed, then translocate to the nucleus and bind their target sites with high specificity. Upon binding the target sites, the FokI nuclease forms its catalytically active dimer and creates a single, specific double-strand break at the user-defined locus. Living cells have evolved several methods to repair double-strand breaks, and these well-studied, endogenous processes can be harnessed to create gene knockouts or knock-ins (Figure 2). Following the creation of a double-strand break, the cell chooses to repair it either via non-homologous end joining (NHEJ) or via homologous recombination (HR). NHEJ is an imperfect repair system, and can lead to insertions or deletions of base pairs at the site of the double-strand break. The mutation leads to the creation of a frameshift within the coding sequence of the gene, an in-frame deletion or exon skipping that may disrupt gene translation and knockout gene function. If the cell is provided with a donor plasmid that has a sequence homologous to regions immediately flanking the double-strand break, the cell may use the donor for homology directed repair or HR. The donor plasmid can be designed to include transgenes for targeted integration, create specific mutations at the endogenous locus, or tag endogenous genes. Here we describe an example of a ZFN-induced gene knockout in a human lung cancer cell line, a ZFN-induced targeted integration of a miRNA at the AAVS1 locus in the human genome, and a ZFN-induced tagging of human tubulin with green fluorescent protein.

Figure 2 The addition of zinc finger nucleases to the cell results in creation of a double-strand break at the target site. This double-strand break is repaired by one of two endogenous repair pathways, either the non-homologous end joining (NHEJ) or the homologous recombination (HR) pathway. NHEJ is used to create gene knockouts while HR is utilized for targeted integration.

Targeted Knockout of BAX in Human Lung Cancer

Previously, mammalian cell line engineering techniques to disrupt specific genes was limited to time consuming homologous recombination and screening approaches. The advent of ZFN technology now enables simple, rapid and permanent disruption of specific gene loci. Sequence specific nuclease cleavage, followed by imperfect DNA repair, gives rise to permanent gene mutations. ZFNs introduced against a particular gene create a single, specific double-strand break at the target genomic locus. This activates the NHEJ repair pathway that is referred to as “nonhomologous” because the break ends are directly ligated without the need for a large homologous template. NHEJ usually repairs the break accurately, but sometimes the broken DNA is acted upon by nucleases and polymerases that can lead to the insertion and/or deletion of bases. This insertion or deletion of bases leads to disruption of the coding sequence and functional loss of the protein. It is this less frequent outcome of incorrect repair that enables the targeted DSBs from the ZFNs to result in the specific loss of gene product and function.

Here, we demonstrate an example of the rapid construction of genetically defined cell lines using ZFNs (The protocol used here is in the technical bulletin for CSTZFN). The pro-apoptotic Bcl-2–associated X protein, BAX, was targeted for deletion in three human cancer cell lines: A549, DLD-1, and SW48. In all cases, even in A549 cells where BAX is tetraploid, knockout lines were successfully generated. An immunometric assay was used to confirm that BAX protein is not produced in the complete knockout cell lines, and protein production was decreased in partial knockout cell lines. In addition to multi-allelic disruptions, cell lines with monoallelic disruptions were also isolated. DNA sequence alterations were verified by sequencing, and the absence of BAX protein was confirmed by enzyme immunometric assay (Figure 3). The straightforward approach coupled with the power of ZFN technology to address both diploid as well as polyploid targets, makes this an ideal tool in the generation of these cell lines. Targeted gene editing provides a critical tool that enables the study of genes involved in cancer and other processes.

Figure 3 Knockout of tetraploid BAX in A549 cells using ZFNs. (A) Following treatment with a ZFN specific for BAX, a clone was isolated that contained a unique disruption in all four alleles of BAX. The DNA sequences of the wildtype and four disrupted alleles are shown above. (B) BAX protein concentration was measured in wild type and knockout cell lines using an enzyme immunometric assay specific for BAX. Quantification of BAX protein levels was obtained through comparison of BAX protein levels in the cell lysate to a standard curve of recombinant BAX protein. Clones having 1, 2 or 3 out of 4 alleles disrupted produce less BAX protein. For three unique clones, each with all 4 alleles disrupted, the measurement was below the lower limit of detection demonstrating that no BAX protein is produced when 4 out of 4 alleles are disrupted. Clones with 1, 2 or 3 of 4 alleles disrupted produce less BAX protein (The linear detection limit for the assay is at 15 pg). Data provided by Suzanne Hibbs and Gregory Wemhoff, Ph.D., at Sigma Life Science.

Targeted Integration of a miRNA at the AAVS1 locus

The targeting power of the ZFN technology allows placement of transgenes precisely into user-specified sites in the genome with high-efficiency. Traditional use of random insertion to introduce a DNA construct into the human genome may lead to gene disruption, variable epigenetic effects on transgene expression and presence of an unknown number of integrants in each cell, making comparisons difficult between isolated single cell clones. ZFN-induced targeted integration of a transgene into a specific locus in the human genome enables stable transgene expression, regulation of copy number and site of integration, without the need to engineer cell lines with randomly integrated transgene landing pads, e.g., the insertion of recombinase sites within the genome. The specificity of targeting comes not only from the homology arms in the donor, but also from the site of the DSB created by site-specific ZFNs. Upon creation of this DSB the cell may use HR for repair using a homologous template, typically the sister chromatid. We usurp this natural process by flooding the cell with a donor DNA template bearing homologous sequences to the site of the genetic lesion and any additional genetic information the user wants to insert at that locus between the homology arms (Figure 4). The cell uses the donor as the repair template and subsequently integrates the desired transgene into the locus. This form of repair can be thought of as a “copy and paste” mechanism having a higher fidelity than NHEJ, but requiring a longer region of homology.

Figure 4 (A) Schematic showing the construction of A549 and MCF7 cells with ZFNs and pZDonor plasmid containing a microRNA (miRNA) gene. The pZDonor plasmid containing the miRNA was co-nucleofected into A549 and MCF7 cells with mRNAs that express AAVS1-specific ZFNs. These ZFNs cut at the AAVS1 locus within the human genome (binding site in red). The miRNA gene is integrated into the genome through homologous recombination using the homology arms on the pZDonor plasmid. The miRNA is driven by a PGK (phosphoglycerate kinase) promoter and has a synthetic poly(A) signal. (B) The miRNA is expressed and processed correctly in A549 and MCF7 cells. Human miRNA-373, miRNA-520c and tandem (cluster) miRNA520c- 373 were cloned into the pZDonor vector. All three combinations were used in A549 cells and only miR-373 was used in MCF7 cells. Values are expressed on a log scale as percent expression compared to the wild type. (C) Endogenous CD44 mRNA levels are reduced in MCF7 cells. CD44 is a known target for human miRNA-373. Endogenous levels of CD44 mRNA were reduced upon expression of miRNA-373. Values are expressed as percent expression compared to the wild type. Data provided by Kevin Forbes, Ph.D., and Carol Kreader, Ph.D., at Sigma Life Science.

Here we demonstrate an example of the use of a safe harbor site, a genomic locus that maintains transcriptional competence and has no known adverse effect on the cell from its disruption, for the expression of miRNA genes in metastatic cancer cells. Earlier studies have demonstrated the function of miRNA genes in metastatic cancer cells using lentiviral introduction of the miRNAs. The lentiviral introduction of the miRNAs will lead to random integration of the construct into the human genome with the risk of deleterious effects on the integration site. The use of a safe harbor locus (such as Rosa26 in mice) will allow expression of a construct from a known genetic and epigenetic locus. We used the AAVS1 site, which has been characterized as a safe harbor locus. We were able to express one or multiple miRNAs from a single locus and successfully demonstrate the knockdown of CD44, a known target of the expressed miRNA. This method demonstrates that the successful use of the AAVS1 safe harbor locus for the creation of transgenic human cell lines expressing miRNA genes as individuals or as clusters. Recent work has also demonstrated that CCR5 can also be used as an additional safe harbor locus to create transgenic human cells.

Tagging Endogenous Tubulin with GFP

As discussed earlier, ZFN technology is a fast and reliable way to manipulate the genome in a targeted fashion. In the majority of cases ZFNs have been used to create gene knockouts utilizing non-homologous end joining (NHEJ) – one of the main pathways of double-strand break (DSB) repair in somatic cells. Here we relied on the second DSB repair pathway – homologous recombination (HR) to tag cytoskeletal genes by integrating a fluorescent reporter sequence into the desired location in the genome. The integration resulted in endogenous expression of the corresponding cytoskeletal protein fused to a fluorescent protein. These fusion proteins have the advantage of easy and quick detection at their native expression level and characteristic localization pattern in the cell. Three gene loci were tagged: LMNB1 (lamin B1, nuclear envelope), TUBA1B (α-tubulin 1b, microtubules: Figure 5), & ACTβ (β-actin, actin fibers). Green and red fluorescent proteins were used as reporters. Post transfection into U2OS cells, fluorescent cells were isolated using fluorescence activated cell sorting (FACS) to identify single cells with gene integration at the endogenous locus. These cells were further characterized for gene expression via southern blotting and western blotting to determine integration specificity and protein expression levels. Moreover, we were also able to isolate cells that had received dual treatment with ZFNs and donor, and were double integrated for fluorescent protein fusions of Lamin B1 and β-Actin. This demonstrates the ability to stack different traits in a single cell line.

Figure 5 U2OS cell lines with TUBA1B tagged with green fluorescent protein (GFP) and red fluorescent protein (RFP) at the endogenous locus. (A) Schematic of the TUBA1B loci and donor including the length of the homology arms, the position of the fluorescent protein at the genomic locus and the preservation of the splice site on the first exon of the gene. (B) DIC & fluorescence microscopy images of endogenously labeled TUBA1B with GFP and RFP in U2OS cells. (C) Southern hybridizations were performed on DNA isolated from wild-type U2OS and nine single cell clones positive for red tubulin fluorescence. Using the tubulin probe, the ~1900bp band represents the RFP-fused tubulin genomic locus while the ~1200bp band represents wildtype (wt) tubulin genomic locus. Targeted integration (TI) in these cells occurred in a heterozygous manner as seen by the presence of both bands in the lanes 1-9. Lane 4 also has a random integration event aside from the targeted integration. Data provided by Hongyi Zhang and Nathan Zenser, Ph.D., and Dmitry Malkov, Ph.D., at Sigma Life Science.

back to top Back to Top

Related Links