Creating Zinc Finger Nucleases to Manipulate the Genome by pengtt


									Creating Zinc Finger Nucleases to Manipulate the Genome in a

Site-Specific Manner Using a Modular-Assembly Approach

                               Matthew Porteus

Adapted from Gene Transfer: Delivery and Expression of DNA and RNA (ed. Friedmann
            and Rossi). CSHL Press, Cold Spring Harbor, NY, USA, 2007.


Homologous recombination is the most precise way to manipulate the

    genome. It has been used extensively in bacteria, yeast, murine

embryonic stem cells, and a few other specialized cell lines, but it has

not been available in other genetic systems such as mammalian somatic

cells. However, the creation of a gene-specific DNA double-strand break

 can stimulate homologous recombination by several-thousandfold in

mammalian somatic cells. These double-strand breaks can be created in

mammalian genomes by zinc finger nucleases (ZFNs), artificial proteins

 in which a zinc finger DNA-binding domain is fused to a nonspecific

nuclease domain. This article describes how to identify potential targets

for ZFN cutting. It also focuses on how to assemble ZFNs to recognize

target sequences of the form 5'-GNNGNNGNN-3'. It is likely that there

     will be improvements in making the individual finger modules.

 High-throughput methods that combine both selection and assembly

are being developed, which will broaden the number of sites that can be
 targeted, optimize the overall structure of the ZFNs, and increase the

         possible experimental and therapeutic uses for ZFNs.

                       RELATED INFORMATION

  Protocols are available for the Analysis of Targeted Chromosomal

   Deletions Induced by Zinc Finger Nucleases (Kim et al. 2010), for

 Creating Zinc Finger Nucleases Using a Modular-Assembly Approach

     (Porteus 2010a) to cut specific target sites, and for Testing a

   Three-Finger Zinc Finger Nuclease Using a GFP Reporter System

 (Porteus 2010b) to determine if the newly designed ZFN is active in a

                mammalian cell-culture-based system.


   In this article, gene targeting is defined as the transfer of genetic

  information from an introduced fragment of DNA to the genome by

  homologous recombination. It represents the most precise way to

 manipulate the genome and can be used to introduce changes both

     small (e.g., single nucleotide changes) and large (such as the

introduction of several thousand base pairs of new DNA) into the genetic

material of a cell. Gene targeting has been a powerful experimental tool

in the study of bacteria, yeast, murine embryonic stem cells (Capecchi
 1989), and certain other specialized vertebrate cell lines such as the

   chicken DT40 cell line (Buerstedde and Takeda 1991). Moreover,

   because of its precision, gene targeting would be an ideal way to

 perform gene therapy. However, the broad experimental use of gene

targeting in vertebrate cells and its use for therapeutic purposes have

  been precluded by its low spontaneous rate. In HEK-293 cells, for

example, the spontaneous rate of gene targeting is approximately one

   event per million cells transfected (Porteus and Baltimore 2003).

A critical finding in the development of homologous recombination as a

tool was the discovery that a DNA double-strand break in the genomic

        target gene could stimulate targeting at that locus by

 several-thousandfold (Jasin 1996). These experiments are generally

 performed by introducing the recognition site for the I-SceI homing

endonuclease (a DNA endonuclease with an 18-bp recognition site) into

   a reporter gene and then introducing the modified reporter as a

  transgene into a cell of interest. The rate of gene targeting is then

  measured by the correction of the mutated reporter gene after the

 introduction of the I-SceI nuclease and a donor plasmid to serve as a

template for homologous recombination. An obvious limitation of this

 strategy is that no human gene has a natural recognition site for the

                            I-SceI nuclease.
Another critical finding was the discovery that ZFNs, previously known

as chimeric nucleases, could stimulate gene targeting in human somatic

cells (Porteus and Baltimore 2003). ZFNs are artificial proteins formed by

   the fusion of a zinc finger DNA-binding domain to a nonspecific

nuclease domain (Kim et al. 1996). Thus far, the only nuclease domain

examined in a ZFN is the one derived from the FokI type-IIS restriction

endonuclease. The ability of ZFNs to stimulate gene targeting was first

 demonstrated using model ZFNs in which the zinc finger domain was

     derived from a zinc finger protein with a known binding site.

   Subsequently, ZFNs were made de novo to recognize novel target

sequences from natural genes (Alwin et al. 2005; Porteus 2006). These

   nucleases could stimulate targeting several-thousandfold using a

    transgenic green fluorescent protein (GFP) reporter assay. More

    significantly, Urnov et al. (2005) demonstrated that ZFNs could

stimulate targeting at the endogenous IL2RG locus in up to 20% of cells.

  These authors designed the IL2RG ZFNs by assembling zinc finger

 domains from a proprietary archive and subsequently optimizing the

zinc fingers by using specialized knowledge of zinc finger-DNA binding.


    Construction of New Zinc Finger Proteins by Modular Assembly
   Zinc finger proteins can be assembled using previously described

 individual zinc fingers. This success of this approach depends on: (1)

the modular nature of zinc finger binding to DNA and (2) the published

data sets of individual zinc fingers and their cognate DNA-binding sites.

 The modular nature of zinc finger binding was revealed by the crystal

structure of the zinc finger domain (Pavletich and Pabo 1991; Pabo et al.

2001). An individual zinc finger consists of 30 amino acids arranged in a

  ββ   structure that is stabilized by chelating a single zinc ion. The

-helix of the zinc finger lies in the major groove. The binding of DNA is

mediated by amino acid residues -1 to 6 of the      -helix (with respect to

                       the beginning of the helix).

A zinc finger protein consists of a series of individual zinc fingers. Thus,

   a three-finger protein is one that has three individual zinc finger

  domains. The modular nature of zinc finger binding has two crucial

   aspects: (1) Each individual finger binds a nonoverlapping triplet

   independently of its neighboring finger and (2) each finger makes

contact with each base of the triplet with a single amino acid. Because of

  these attributes, it is possible to create a new finger that binds to a

different 3-bp sequence by altering the amino acid contact residues of

an individual finger. Furthermore, by shuffling different individual zinc

fingers, one can create a new zinc finger protein with a new target site

 specificity. For a three-finger protein, for example, that would mean
creating a protein with a novel 9-bp binding site. The testing, successes,

    and limitations of the modular model of zinc finger binding are

 described elsewhere (Wolfe et al. 2000; Pabo et al. 2001). A variety of

      individual fingers have been published that bind to unique

triplet-binding sites, including triplets of the form 5'-GNN-3' (Segal et

 al. 1999; Dreier et al. 2000; Liu et al. 2002), 5'-ANN-3' (Dreier et al.

               2001), and 5'-CNN-3' (Dreier et al. 2005).

Using published data sets, a number of different artificial transcription

factors have been made (Segal et al. 2003). Almost all of them have been

 designed to recognize target sequences rich in 5'-GNN-3' target site

   triplets. The reason for this bias could be because the published

individual zinc fingers that recognize GNN triplets are of higher quality

 than those that bind to non-GNN triplets. On the other hand, careful

  studies of zinc finger DNA binding have shown that binding is not

 completely modular and that binding of a given finger depends on its

neighbors, i.e., there is "context" dependence. An important caveat to

the modular-assembly approach for designing new zinc finger proteins

is that this context dependence is ignored. Thus, fingers that bind GNN

triplets might be more modular in their binding and depend less on the

          context of the binding of the neighboring domains.

                   Identification of ZFN Target Sites
  To cut DNA, the FokI nuclease domain must dimerize. Dimerization

occurs when two ZFNs bind to their cognate binding sites in the correct

 orientation. Each ZFN binds to its own 9-bp recognition site. A single

ZFN-binding site is referred to here as a "ZFN half-site" because a single

 site is not sufficient for cleavage. A pair of ZFN half-sites oriented in

inverse orientation (necessary for nuclease dimerization) is called a "ZFN

full-site." The spacer between the two ZFN half-sites should be 5 or 6 bp.

  There is evidence from mammalian systems and others that longer

  spacers are less effective, although studies of shorter spacers and

systematic comparison of different spacer lengths in mammalian cells

have not been performed. Sites are referred to as a "ZFN full-site(5)" or a

   "ZFN full-site(6)", depending on whether the spacer between two

                   half-sites is 5 or 6 bp, respectively.

The first step to designing a pair of ZFNs is to identify a ZFN full-site in

the target gene of interest. Although an active ZFN that recognizes the

 target site 5'-AGGGATAAC-3' has been prepared (Alwin et al. 2005),

published zinc finger domains that recognize 5'-GNN-3' seem to work

best for modular assembly. Future work will determine how often ZFNs

can be made using the modular-assembly approach to recognize target

sequences that do not contain all 5'-GNN-3' triplets (Alwin et al. 2005),

 but for the purposes of this article, searches were limited to half-site

 targets of the form 5'-GNNGNNGNN-3'. The program DNAsis can be
             used to identify sequences either of the form

     5'-NNCNNCNNCnnnnnnGNNGNNGNN-3' (ZFN full-site[6]) or

5'-NNCNNCNNCnnnnnGNNGNNGNN-3' (ZFN full-site[5]); other search

    programs can be used to identify possible ZFN full-sites. In the

   GFP-coding region, there are three ZFN full-sites(6) and two ZFN

 full-sites(5) and ZFNs have been assembled to target the most 5' ZFN

                       full-site(6) (Porteus 2006).

After identifying a potential ZFN full-site, the next step is to design the

  zinc finger proteins to recognize each ZFN half-site. There are two

major published data sets for zinc finger domains that recognize GNN

triplets (Segal et al. 1999; Liu et al. 2002), although they have not been

 compared formally to determine which might be better. Both Porteus

(2006) and Alwin et al. (2005) attempted to make ZFNs to recognize two

ZFN half-sites in the GFP gene. Porteus was successful in making active

ZFNs to both half-sites using the data set of Liu et al. (2002), whereas

Alwin et al. (2005) were only successful at one of the two using the data

 set of Segal et al. (1999). The data set of Liu et al. (2002) has proved

  successful, albeit with varying degrees of efficiency, in 10 out of 10

attempts at targeting ZFN half-sites of the form 5'-GNNGNNGNN-3', as

measured by the GFP gene-targeting reporter system described below.
          One of the ZFN full-site(6) targets in the GFP gene is

  5'-ACCATCTTCttcaagGACGACGGC-3'. Therefore, one ZFN must be

designed to recognize the ZFN half-site(1) 5'-GAAGATGGT-3' and the

other to recognize the ZFN half-site(2) 5'-GACGACGGC-3'. Zinc fingers

  bind to DNA in an antiparallel fashion such that finger 1 (the most

 amino-terminal finger) binds to the most 3' triplet (Durai et al. 2005;

 Porteus and Carroll 2005). Thus, finger 1 of GFP1-ZFN must bind to

5'-GGT-3', finger 2 to 5'-GAT-3', and finger 3 to 5'-GAA-3'. The amino

 acid composition of the recognition      -helix is then deduced by using

Figure 2 of Liu et al. (2002). For GFP1-ZFN, this results in the -helix of

 finger 1 being QSSHLTR, finger 2 being TSGNLVR, and finger 3 being

  QSGNLAR (using the standard one-letter amino acid abbreviations).

  Table 1 of Porteus (2006) describes the amino acid content of the

  -helix with the desired DNA target binding site for five other ZFNs,

including GFP2-ZFN. Table 2 of Liu et al. (2002) shows that the order of

the fingers is reversed from the order they occur in the designed ZFN.

For example, because finger 1 of GFP1-ZFN is designed to recognize the

triplet 5'-GGT-3' (i.e., the 3'-most triplet of the overall binding site), the

  amino acids from "Finger 3" in Table 2 of Liu et al. (2002) are used

            because that is the 3'-most triplet in their table.

          Using PCR to Assemble a New Three-Finger Protein
  An overlapping polymerase chain reaction (PCR) strategy is used to

assemble the new zinc finger proteins (Fig. 1 ; Segal et al. 2003). Each

 finger is amplified independently using a general primer at the 5' end

and a finger-specific primer at the 3' end. The Zif268 backbone is used

as a template for each PCR because there is enough heterogeneity in the

backbone to allow assembly of the three fingers in the correct order in

the final step. In some artificial ZFN constructs, such as QQR-ZFN, the

nucleotide sequence surrounding the recognition helix of each finger is

 identical, which makes it very difficult to assemble the fingers in the

correct orientation by overlap PCR in the next step. Each finger fragment

is amplified such that it has a 15-bp overlap with its neighboring finger

  and the individual fingers are then assembled using an overlap PCR

 strategy. The final PCR product is digested with BamHI and SpeI and

cloned into BamHI/SpeI-digested pBluescript. The three-finger cassette

  is then sequenced to verify that no errors were created during the

   process. Although a high-fidelity polymerase was used initially,

subsequent studies found that standard Taq was accurate enough and

    did not create undesired point mutations. The sequences of the

  oligonucleotides used in this process are shown in Table 1 and the

    procedure is described in greater detail in Creating Zinc Finger

   Nucleases Using a Modular-Assembly Approach (Porteus 2010a).
                                Figure 1. Schematic overview of assembly of

                                new zinc finger proteins. The experimental

                                details of the procedure are outlined in the

                                text. The colors represent the unique amino
   View larger version (23K):
                                acids of the zinc finger   -helix that mediate
        [in this window]
                                specific DNA binding for each finger.
      [in a new window]

                     Testing a New Three-Finger ZFN

There are a number of ways to test the activity of a newly designed ZFN.

Protein can be made via in vitro transcription/translation or from some

other expression system, and the ability to cut a DNA template in vitro

  can be tested. Originally, designed ZFNs were evaluated by in vitro

  cutting using bacterially expressed and purified proteins (Kim et al.

    1996; Smith et al. 1999, 2000). Alwin et al. (2005) described a

cell-based transcription assay to test the new zinc finger portion of the

   ZFN for binding to its cognate binding site. Our laboratory uses a

GFP-based reporter system (Fig. 2 ; Porteus and Baltimore 2003; Porteus

  and Carroll 2005). The method is discussed more fully in Testing a
Three-Finger Zinc Finger Nuclease Using a GFP Reporter System

                             (Porteus 2010b).

                             Figure 2. Schematic of GFP gene-targeting

                             reporter system. To measure gene targeting, a

                             cell line is created with a single copy of the GFP

                             reporter gene. The GFP reporter gene contains

                             recognition sites for the I-SceI endonuclease,

View larger version (33K):   the GFP-ZFNs, and a full-target site created

     [in this window]        from the Zif268 half-site and the target site for

   [in a new window]         the newly assembled ZFN. Nuclease expression

                             constructs are then transfected into the reporter

                             line along with a repair donor plasmid using

                             standard techniques. Three days after

                             transfection, the cells are analyzed by flow

                             cytometry for the number of GFP-positive cells.

                             CMV/CBA, hybrid cytomegalovirus/chicken

                             β-actin enhancer promoter; IRES, internal

                             ribosomal entry site; WRE, woodchuck

                             post-transcriptional regulatory element.

M.P. thanks the Burroughs-Wellcome fund and the University of Texas

 Southwestern Medical Center for their support of the research in his

  laboratory. The work in M.P.’s lab was also supported by National

                  Institutes of Health grant K08 HL70268.


       1.   Alwin S, Gere MB, Guhl E, Effertz K, Barbas CF III, Segal DJ,

       Weitzman MD, Cathomen T. 2005. Custom zinc-finger nucleases

            for use in human cells. Mol Ther 12: 610–617.[Medline]

  2.    Buerstedde J-M, Takeda S. 1991. Increased ratio of targeted to

       random integration after transfection of chicken B cell lines. Cell

                              67: 179–188.[Medline]

       3.   Capecchi MR. 1989. Altering the genome by homologous

       recombination. Science 244: 1288–1292.[Abstract/Free Full Text]

 4.    Dreier B, Segal DJ, Barbas CF III. 2000. Insights into the molecular

        recognition of the 5'-GNN-3' family of DNA sequences by zinc

               finger domains. J Mol Biol 303: 489–502.[Medline]
     5.     Dreier B, Beerli RR, Segal DJ, Flippin JD, Barbas CF III. 2001.

            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 276:

                        29466–29478.[Abstract/Free Full Text]

     6.    Dreier B, Fuller RP, Segal DJ, Lund CV, Blancafort P, Huber A,

     Koksch B, Barbas CF III. 2005. Development of zinc finger domains

      for recognition of the 5'-CNN-3' family DNA sequences and their

          use in the construction of artificial transcription factors. J Biol

                  Chem 280: 35588–35597.[Abstract/Free Full Text]

7.    Durai S, Mani M, Kandavelou K, Wu J, Porteus MH, Chandrasegaran

           S. 2005. Zinc finger nucleases: Custom-designed molecular

      scissors for genome engineering of plant and mammalian cells.

            Nucleic Acids Res 33: 5978–5990.[Abstract/Free Full Text]

           8.    Jasin M. 1996. Genetic manipulation of genomes with

     rare-cutting endonucleases. Trends Genet 12: 224–228.[Medline]

     9.     Kim YG, Cha J, Chandrasegaran S. 1996. Hybrid restriction

      enzymes: Zinc finger fusions to Fok I cleavage domain. Proc Natl

                   Acad Sci 93: 1156–1160.[Abstract/Free Full Text]

      10. Kim S, Lee HJ, Kim E, Kim J-S. 2010. Analysis of targeted

          chromosomal deletions induced by zinc finger nucleases. Cold
                        Spring Harb Protoc doi:

            10.1101/pdb.prot5477.[Abstract/Free Full Text]

11. Liu Q, Xia Z, Case CC. 2002. Validated zinc finger protein designs

          for all 16 GNN DNA triplet targets. J Biol Chem 277:

                  3850–3856.[Abstract/Free Full Text]

12. Pabo CO, Peisach E, Grant RA. 2001. Design and selection of novel

          Cys2His2 zinc finger proteins. Annu Rev Biochem 70:


   13. Pavletich NP, Pabo CO. 1991. Zinc finger-DNA recognition:

   Crystal structure of a Zif268-DNA complex at 2.1 Å. Science 252:

                   809–817.[Abstract/Free Full Text]

14. Porteus MH. 2006. Mammalian gene targeting with designed zinc

           finger nucleases. Mol Ther 13: 438–446.[Medline]

    15. Porteus M. 2010a. Creating zinc finger nucleases using a

      modular-assembly approach. Cold Spring Harb Protoc doi:

            10.1101/pdb.prot5530.[Abstract/Free Full Text]

 16. Porteus M. 2010b. Testing a three-finger zinc finger nuclease

       using a GFP reporter system. Cold Spring Harb Protoc doi:

            10.1101/pdb.prot5531.[Abstract/Free Full Text]

 17. Porteus MH, Baltimore D. 2003. Chimeric nucleases stimulate

    gene targeting in human cells. Science 300: 763.[Free Full Text]
 18. Porteus MH, Carroll D. 2005. Gene targeting using zinc finger

           nucleases. Nat Biotechnol 23: 967–973.[Medline]

    19. Segal DJ, Dreier B, Beerli RR, Barbas CF III. 1999. 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 96:

                  2758–2763.[Abstract/Free Full Text]

  20. Segal DJ, Beerli RR, Blancafort P, Dreier B, Effertz K, Huber A,

   Koksch B, Lund CV, Magnenat L, Valente D et al. 2003. Evaluation

   of a modular strategy for the construction of novel polydactyl zinc

             finger DNA-binding proteins. Biochemistry 42:


21. Smith J, Berg JM, Chandrasegaran S. 1999. A detailed study of the

     substrate specificity of a chimeric restriction enzyme. Nucleic

            Acids Res 27: 674–681.[Abstract/Free Full Text]

 22. Smith J, Bibikova M, Whitby FG, Reddy AR, Chandrasegaran S,

     Carroll D. 2000. Requirements for double-strand cleavage by

    chimeric restriction enzymes with zinc finger DNA-recognition

                     domains. Nucleic Acids Res 28:

                  3361–3369.[Abstract/Free Full Text]
23. Urnov FD, Miller JC, Lee Y-L, Beausejour CM, Rock JM, Augustus S,

   Jamieson AC, Porteus MH, Gregory PD, Holmes MC. 2005. Highly

     efficient endogenous human gene correction using designed

         zinc-finger nucleases. Nature 435: 646–651.[Medline]

  24. Wolfe SA, Nekludova L, Pabo CO. 2000. DNA recognition by

   Cys2His2 zinc finger proteins. Annu Rev Biophys Biomol Struct 29:


To top