An improved zinc-finger nuclease architecture for highly specific genome editing

From Wiki FKKT

Jump to: navigation, search

(Eva Knapič)

Miller, J.C.; Holmes, M.C.; Wang, J.; Guschin, D.Y.; Lee, Y.L.; Rupniewski, I.; Beausejour, C.M.; Waite, A.J.; Wang, N.S.; Kim, K.A.; Gregory, P.D.; Pabo, C.O.; Rebar, E.J.An improved zinc-finger nuclease architecture for highly specific genome editing. Nat Biotechnol., 2007, 25(7), 778-785.


Genome editing

Genome editing is a type of genetic engineering in which DNA is inserted, removed or replaced from genome using artificially designed nucleases, enzymes that cleave the phosphodiester bonds between nucleotides. The principle of method is based on nuclease cleavage at a desired site of the genome in order to create specific double-stranded break, then put to use the cell’s endogenous mechanisms to repair the induced break by homology-directed repair or non-homologous end joining. There are a few different classes of engineered nucleases that are currently being used: zinc-finger nucleases, transcription activator-like effector nucleases and the CRISPR/Cas system. In this seminar we will focus on zinc finger nucleases. Firstly, we will take a look at basics of zinc-finger nucleases and their mechanism. Then we will discuss improved zinc-finger nuclease architecture for more specific efficiency described by Miller and colleagues in their paper published in Nature Biotechnology in 2007[1].

Zinc-finger nuclease

Zinc-finger nuclease (ZFN) is an artificially designed endonuclease that can be customised to cleave at sequence specific site on the DNA. The enzyme consists of two domains: DNA-binding domain composed of zinc finger structural motifs and DNA-cleavage domain of restriction enzyme FokI. Domains are connected with short inter-domain linker. This structure combines key qualities of DNA binding specificity, flexibility of zinc-finger motifs and cleavage activity of FokI catalytic domain that is robust but restricted. Furthermore all three elements can be optimised for retargeting and improving efficiency[1][2].

DNA-binding domain

As mentioned above, zinc finger motifs represent DNA-binding domain of ZFN. Zinc fingers are small protein structural motifs distinguished by coordinated zinc ions that stabilise the protein fold. Each zinc-finger recognises 3 base pairs (bp) of DNA. Contacts between the zinc fingers and DNA are made by α-helix that binds in the DNA major groove[3]. Proteins that contain any number of zinc finger motifs are called zinc-finger proteins (ZFPs). Engineered ZFPs consist of three to six zinc finger motifs thus enabling binding of 9 bp to 18 bp targets. Longer recognition sequence improves specificity and precision of ZFPs. ZFP based DNA-binding domains can be coupled to various effector domains, which then cuts the DNA sequence determined by the ZFP[1][2].

There are multiple approaches to designing new ZFPs with binding specificities chosen by user. The simplest and widely used method to generate new zinc finger sets is modular assembly. Candidate ZFPs for target sequence are obtained by determining individual zinc fingers that bind each triplet and assembling them together thus recognising the target sequence[2]. Alternative methods for designing new ZFPs include two-finger modules, oligomerised pool engineering (OPEN) and Context-Dependent Assembly (CoDA). In two-finger modules instead of individual fingers, construct of two-fingers is used for recognising target DNA sequence. Advantage of this approach is better optimisation of finger junctions and speed of finding and assembling suitable zinc fingers, a limitation is extension of initial characterisation of two-finger units [4]. The OPEN system approach is carried out in two steps, firstly appropriate finger pools are recombined to create small combinatorial library, then members of library that efficiently bind to the target site are isolated using bacterial two-hybrid selection method in which binding of a zinc finger domain to its target activates expression of selected marker gene[5]. CoDA is publicly available platform of reagents and software. With this approach three-finger sets are assembled using N- and C- terminal fingers that have been previously identified in other arrays with common middle finger. CoDA does not treat fingers as independent modules but instead explicitly accounts for context-dependent effects between adjacent fingers, increasing the probability of efficiency [6]. Regardless of which approach is used for designing new ZFPs in vitro, their application in living cells is not always successful. Reason is the complexity of genome that often contains naturally occurring multiple copies of sequence that are either identical or very similar to target sequence and these copies can act as additional targets for ZFNs. Another complication is chromatin structure at target sites that may obstruct cleavage[4].

Inter-domain linker

Inter-domain linker connects DNA-binding domain with DNA-cleavage domain. Length of the linker is responsible for the shaping of spacer. Spacer is short sequence between both recognition sites for DNA-binding domains, which enables precise definition of cleavage site. Wilson and colleagues have demonstrated that sites with 5 bp spacer lengths showed the most efficient ZFN mediated gene targeting when the linker is 2 or 4 amino acids and that sites with 3 or 4 bp spacers could not be targeted efficiently. Group also demonstrated that ZFN variants with 5 amino acids linker can efficiently target sites with a 7 bp spacer. However, the ideal linker with corresponding spacer can only be determined by taking into consideration both ZFN activity on the chromosomal target site and ZFN associated toxicity [7].

DNA-cleavage domain

DNA-cleavage domain of ZFN originates from FokI catalytic domain. FokI is type IIS restriction endonuclease consisting of an N-terminal DNA-binding domain and non-specific DNA-cleavage C-terminal domain. In order to cleave DNA two FokI catalytic domains must form dimer. Each subunit contains one catalytic centre that is responsible for cleaving one strand of DNA duplex. This characteristic of FokI catalytic domain is crucial for successful genome editing with ZFN. Dimerization requires two independent binding events, which must occur in correct orientation and spacing between monomers. DNA-cleavage domain is usually fused to the C-terminus of DNA-binding domain, thus two individual ZFNs must bind to opposite strands of DNA. This enables specific targeting of long and potentially unique recognition sites[2].

ZFN-mediated genome editing

ZFNs create double-stranded breaks (DSBs) in DNA at sequence specific site. DSBs can be repaired by cell’s endogenous mechanisms: homology-directed repair (HDR) or non-homologous end joining (NHEJ). There are few different possible outcomes of DSB repair. If the break is repaired by NHEJ, it can lead to gene disruption, tag ligation or large deletion. Gene disruption occurs when two broken ends are ligated back together that can lead to small insertions or deletions at the break site causing target gene disruption. Tag ligation occurs when adaptor with sticky ends is provided and is ligated into the chromosome producing tagged allele. When two different ZFNs cleave simultaneous same chromosome, entire segment between them is deleted. If the break is repaired by HDR, which occurs in the presence of donor DNA, it can result in gene correction, targeted gene addition or transgene stacking. If donor with small change, for example single bp, is provided, editing of endogenous allele results in gene correction. This enables easier functional genomics studies, modelling of disease caused by point mutation and gene therapy research. If provided donor carries transgene and has specific homology arms, transgene can be efficiently integrated into chromosome at break site, which results in targeted gene addition. Transgene stacking occurs when provided donor carries multiple linked transgenes between the homology arms. With this approach not only gene addition can be done, but also gene correction if DSB is created within or near gene in question and provided donor carries modified gene sequence[2][4][8].


Model organisms

ZFNs enable introduction of targeted modification in several model organisms used in common biological researches. Benefits of DSB repaired by cell’s endogenous mechanisms include: simple production of knock-out or knock-in cell lines, permanent and heritable mutation, compatibility with variety of species[8].

Therapeutic application of ZFNs

ZFN-mediated gene modification by HDR has been used to correct mutations associated with X-linked severe combined immune deficiency (SCID), haemophilia B, sickle-cell disease and α1-antitrypsin deficiency. ZFN induced gene targeting by NHEJ repair has been shown as potential treatment for HIV/AIDS, either by targeting cellular co-receptors for HIV in primary T cells and hematopoietic stem/progenitor cells (clinical trials) or by targeting proviral HIV DNA[8].

Potential problems

If the DNA-binding domain is not specific enough or target site is not unique in genome, ZFN may cleave at undesired location in the genome. Another potential problem is formation of ZFN homodimers that are by products of ZFN heterodimer that can cleave off-target. Besides cleaving off-target, ZFN homodimers mediate toxic effect. Undesired cleavage can lead to unwanted insertions or deletions at additional cleavage sites or vast number of DSBs can cause cell death. This can be avoided by extensive bioinformatics research beforehand, construction of ZFP that recognises longer and consequently rarer target sequences, using suitable linkers and cleavage domains that do not form unwanted homodimers[1][2].


Miller and colleagues present in their paper an improved ZFN architecture for more specific genome targeting. Their aim was to develop FokI cleavage domain variants that function as obligate heterodimers[1].

Development of heterodimeric FokI cleavage domain variants

There are few things that need to be put in consideration when altering architecture of FokI DNA-cleavage domain. Firstly, FokI is an enzyme, this means goal is to improve interaction between domains but at the same time, catalytic function of domains must be conserved. Secondly, interaction between FokI DNA-cleavage domains in dimer has very low affinity, which is essential for ensuring cleavage specificity. Lastly, interface of FokI dimer is hydrophilic. Considering all the facts, Miller and colleagues developed strategy that involved step by step approach to modification of dimer interface coupled with direct testing of candidate variants for catalytic activity. Reporter system with green fluorescent protein (eGFP) was used to detect gene correction activity. The system uses HEK293 cell line containing a copy of the eGFP gene that is disrupted by short DNA fragment and has characterised ZFN target sequence (Figure 2a). Screening is based on ZFN cleavage of DNA fragment that is corrected by HDR using exogenous donor DNA containing missing eGFP sequence as template. If repair is successful, green fluorescence can be detected by fluorescence-activated cell sorting (FACS)[1]. FACS is specialized type of flow cytometry, method for separating heterogeneous mixture of biological cells, which is based upon the specific light scattering at fluorescent wavelength[9]. Results are shown in Figure 2b, left for cells transfected with donor DNA only and right for cells transfected with donor DNA and ZFNs. Step by step modification of interaction surface was carried out in four cycles. Modifications were implemented alternately, which means in each cycle only one partner domain was modified. Mutations were designed to form charge-charge interaction. Each variant was tested for gene correction activity as a heterodimer with unmodified partner ZFN and as a homodimer. More in detailed view of introduced mutation can be found in Supplementary table 1. In each cycle researchers identified a variant of domain that induced gene correction as a heterodimer and at the same time the domain had reduced activity as a homodimer (Figure 2d). Variants with best gene correction rate contained double mutations E490K:I538K and Q486E:I499L (also known as EL:KK). In one domain glutamic acid (E) at position 490 and isoleucine (I) at position 538 were replaced with lysine (K) (E490K:I538K), in partner domain glutamine (Q) at position 486 was replaced with glutamic acid and isoleucine at position 499 was replaced with leucine (L) (Q486E:I499L). As lysine is positively charged amino acid, variants with two additional lysine residues were referred to as ‘+’ and the other variants were referred to as ‘-‘, due to negative charge of introduced glutamic acid. Model of how variants may interact with each other is shown in Figure 2e. Variants ‘+’ and ‘-‘ exhibited strong preference for heterodimerization and very weak preference for homodimerization when connected with suitable ZFP (L or R ZFP) in GFP reporter system. In Figure 2f we can see that double mutations heterodimers (marked as L+/R- and L-/R+) have higher efficiency for gene correction than wild-type (wt) dimer (marked as Lwt/Rwt) and homodimers of double mutation variants (marked as L+/R+ and L-/R-) have almost none. Results of GFP reporter system indicate that formation of obligated heterodimer was achieved[1].

Validation of heterodimeric FokI cleavage domain variants

Efficient endogenous gene editing

To confirm that the heterodimer variants preserved catalytic function researchers fused them with ZFPs that target a sequence present in exon 5 of the human gene for interleukin-2 receptor γ-chain (IL2Rγ). Designed ZFNs induced gene editing that introduced novel BsrBI restriction site in IL2Rγ (Figure 3a). Restriction enzyme BsrBI is able to cut at such restriction site resulting in additional band on electrophoresis. Results of gene editing efficiency shown in Figure 3b indicate BsrBI restriction site was successfully inserted in case of ZFN with wt cleavage domain (Lwt/Rwt) and both heterodimeric combinations (L+/R- and L-/R+). Forced homodimers (L+/R+ and L-/R-) yielded no detectable gene editing. These results confirmed obligated heterodimerization of developed variants[1].

Suppression of homodimer function in vitro

Cleavage activity in vitro was directly measured by radiolabeling of target sequences for ZFNs (Figure 4a). Target sequences were then digested by ZFNs with wt cleavage domain and variants of domain with double mutations, both in heterodimeric and homodimeric form. Migration of cleaved and uncleaved products in Figure 4b indicate that wt cleavage domain (Lwt/Rwt) was active in heterodimeric and homomeric form, while efficient cleavage with double mutation variants (L+/R- and L-/R+) was only observed if domains formed heterodimer. This result once again supports development of variants, which form obligated heterodimer[1].

Reduced levels of genome-wide cleavage

The aim of study was to develop more specific ZFNs that do not cleave off-target. Therefore, researchers conducted the next experiment on ZFN specificity of gene modification in mammalian cells. ZFN cleavage was determined by detection of proteins that localise at DNA DSBs. The proteins bound were detected by antibody-mediated technique. Protein that localise at DNA damage site and form foci, is tumour suppressor p53-binding protein 1 (53BP1). Protein 53BP1 was detected by immunofluorescence in target cells, which were transfected with ZFN expression constructs and treated with anti-53BP1 rabbit polyclonal antibodies. Results in Figure 5a show that signal is weaker in cells transfected with ZFN with double mutations variants compared to ZFN with wt cleavage domain. This indicates cleavage activity was reduced in case of transfection with modified ZFN. Western blot analysis (Figure 5b) confirmed that both ZFNs were expressed at comparable levels, suggesting that the reduced activity was a result of decreased off-target effect. Alternative damage marker is phosphorylated histone H2AX (γH2AX), which also forms foci at DSBs. Flow cytometry data (Figure 6a) for target cells transfected with ZFNs and stained with antibodies against γH2AX showed that number of positive cells (cells forming γH2AX foci) was significantly lower for ZFNs with double mutations variants compared to ZFN with wt cleavage domain. In addition gene modification was monitored by Surveyor Nuclease assay, which determines the frequency of the small insertions and deletions (indels) characteristic of unspecific DSB repair by NHEJ. Results (Figure 6b) showed comparable levels of gene modification in cells treated with ZFNs with wt and double mutations cleavage domains, indicating comparable cellular activities at target locus. These results confirmed that heterodimeric FokI cleavage domain variants retain catalytic activity while exhibiting reduced levels of off-target cleavage[1].


In conclusion, Miller and colleagues developed two complementary FokI cleavage domain variants with double mutation, which function as an obligate heterodimer and improve ZFN specificity limiting off-target effect. Researchers predicted that their new architecture of FokI cleavage domain could be beneficial for therapeutic application of ZFNs due to additional mechanism for reducing off-target cleavage. It could also be helpful for less safety-intensive application such as crop engineering, cell-line engineering and construction of disease models. These ZFNs would be useful for gene modification protocols requiring simultaneous cleavage at multiple targets, where it would eliminate unwanted combination of ZFNs[1].

Further research

In 2011 same research group form Sangamo BioSciences published a new paper in Nature Methods discussing improved obligate heterodimeric architectures. In this study yeast-based selection system was used to cross-examine the dimer interface. A reporter assay was used to isolate FokI with mutations causing cold-sensitivity, phenotype resulting in diminished cleavage activity at lower temperatures but still be active at higher ones. The hypothesis of study was that cold-sensitive mutants would affect critical residues involved in dimerization, because this class of mutations is often associated with incorrect assembly of multisubunit protein complexes[10].

A selection system in Saccharomyces cerecisiae developed to isolate ZFN mutants with cold-sensitivity phenotype included two independent single-stranded annealing reporter constructs that were integrated into the yeast genome. DNA sequence for reporter genes MEL1 and PHO5 were interrupted by binding site for ZFN homodimer (CCR5-L) (Figure 1a). Additionally PHO5 reporter system had incorporated positive selection cassette natMX that enables resistance to antibiotic nourseothricin. MEL1 reporter system contained the URA3 gene for negative selection using 5-fluoroorotic acid, which can be converted into the toxic compound causing cell death and kanMX cassette that enables resistance to antibiotic geneticin. ZFN-induced DSB would result in restoration of reporter gene expression and simultaneous elimination of all positive and negative selection markers. Library of FokI mutants was constructed by error-prone PCR that randomly mutagenized the sequence (Figure 1b). ZFN expression was induced at 22 °C, and then cells were collected and incubated in medium with geneticin and nourseothricin. With this step all cells carrying active ZFN constructs were eliminated. Remaining cells were shifted to 37 °C and plated on medium with 5-fluoroorotic acid and colorimetric substrate for Pho5. This step gave them conditionally active ZFNs. Constructs were then isolated and tested at 22 °C, 30 °C and 37 °C to confirm cold-sensitive phenotype (Figure 1c); 16 mutants had minimal activity at 22 °C but had restored reporter gene expression at higher temperature. Within isolated mutants previously altered residues that contributed to obligated heterodimerization (glutamine 486 and isoleucine 499 and 538) were found. Their focus was residues asparagine 496 and histidine 537, which face each other in dimerization interface. These two residues were substituted with a pair of oppositely charged amino acids into EL:KK FokI backbone. This substitution could strengthen dimerization in the context of the obligate heterodimer variants. The asparagine 496 was replaced with negatively charged glutamic and aspartic acid in the EL domain, and histidine 537 was replaced with positively charged lysine and arginine in the KK domain. Measured cleavage activity comparted to both wt and EL:KK versions suggested that aspartic acid 496 drove maximal activity in EL monomer, whereas lysine and arginine at position 537 resulted in similar activity improvements. This domains were referred to as ELD:KKK or ELD:KKR. Improved activity of new FokI variants was confirmed by various assays. All results supported superior cleavage activity and retainment of obligate heterodimer function. Researchers concluded that these enhanced FokI domains were portable to many ZFPs, independent of cell type and are a general solution for improved ZFN activity[10].


  1. 1.00 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 1.10 Miller, J.C.; Holmes, M.C.; Wang, J.; Guschin, D.Y.; Lee, Y.L.; Rupniewski, I.; Beausejour, C.M.; Waite, A.J.; Wang, N.S.; Kim, K.A.; Gregory, P.D.; Pabo, C.O.; Rebar, E.J. An improved zinc-finger nuclease architecture for highly specific genome editing. Nat Biotechnol., 2007, 25(7), 778-785.
  2. 2.0 2.1 2.2 2.3 2.4 2.5 Urnov, F. D.; Rebar, E.J.; Holmes, M. C.; Zhang, H. S.; Gregory, P.D. Genome editing with engineered zinc finger nucleases. Nat Rev Genet., 2010, 11(9), 636-646.
  3. Zinc finger ; Wikipedia the free encylopedia [cited 2.1.2015].
  4. 4.0 4.1 4.2 Carroll, D. Genome engineering with zinc-finger nucleases. Genetics., 2011, 188(4), 773-782.
  5. Maeder M. L. et al. Rapid “open-source” engineering of customized zinc-finger nucleases for highly efficient gene modification. Mol Cell., 2008, 31(2), 294–301.
  6. Sander J.D. et al. Selection-free zinc-finger-nuclease engineering by context-dependent assembly (CoDA). Nat Methods., 2011, 8(1), 67-69.
  7. Wilson, K.A.; McEwen, A.E.; Pruett-Miller, S.M.; Zhang, J.; Kildebeck, E.J.;, Porteus, M.H. Expanding the Repertoire of Target Sites for Zinc Finger Nuclease-mediated Genome Modification. Mol Ther Nucleic Acids., 2013, 2, e88.
  8. 8.0 8.1 8.2 Gaj, T.; Gersbach, C.A.; Barbas, C.F. 3rd. ZFN, TALEN, and CRISPR/Cas-based methods for genome engineering. Trends Biotechnol., 2013, 31(7), 397-405.
  9. Flow cytometry; Wikipedia the free encylopedia [cited 2.1.2015].
  10. 10.0 10.1 Doyon, Y.; Vo, T.D.; Mendel, M.C.; Greenberg, S.G.; Wang, J.; Xia, D.F.; Miller, J.C.; Urnov, F.D.; Gregory, P.D.; Holmes, M.C. Enhancing zinc-finger-nuclease activity with improved obligate heterodimeric architectures. Nat Methods., 2011, 8(1), 74-79.

SB students resources

Personal tools