36 3 Site-Directed Genome Modification with Engineered Zinc Finger Proteins
3.2 Approaches for Engineering or Acquiring Zinc
Finger Proteins
The most common approach for targeted genomic DNA cleavage via ZFPs is to
use ZFNs [19–21]. Simplistically, a ZFN involves fusion of a ZFP to a nuclease
domain via a short flexible linker sequence. The simplest ZFN combines a natu-
rally occurring ZFP with the linker and FokI nuclease domains to target its native
binding site [17]. However, ZFNs can also be rationally designed to target a wide
range of sequences for a greater number of applications.
ZF motifs are 30-amino-acid protein domains that chelate a zinc ion. They
bind to DNA by insertion of an alpha helix into the major groove of the DNA
to probe the DNA sequence [22]. Naturally occurring ZFs may be mutated to
alter binding specificity [23]. The DNA sequence that will be bound is defined
by certain amino acids [24, 25]. These ZFs can be combined into strings of 3,
4, 5, or 6 ZFs to bind increasingly long DNA sequences to enhance the speci-
ficity of the interaction [26–28]. The availability of motifs that recognize tri-
plet sequences is a limiting factor in ZF design, as the ZFN pair should be
designed such that they will create a DSB as close to the site of desired HR as
possible [29].
Most restriction enzymes cleave palindromic sequences through coupled
DNA-binding and cleavage events. The FokI endonuclease is different in that it
cuts DNA between two binding sites that can be 9–18 bp in length on opposite
DNA strands. FokI contains two separate domains: the N-terminal domain is
involved in sequence recognition, while the C-terminal domain contains a nucle-
ase [30, 31]. FokI is unique in that single amino acid substitutions resulted in the
decoupling of sequence recognition and cleavage [32, 33], allowing the nuclease
domain to be isolated and fused to other DNA-binding domains. In addition,
dimerization of the nuclease domain is required for DNA cleavage to occur [34].
The ZFN architecture has been improved such that cleavage by the enzyme
requires a heterodimer to be formed, preventing the off-target events that could
result from homodimer formation [35]. A recent study reported a multi-reporter
selection system to identify ZFNs with high degrees of activity at the desired site
and negligible activity at similar off-target sites in the genome [36]. Refinements
through mutagenesis and DNA shuffling have made the FokI cleavage domain
15-fold more active and 6-fold more specific [37]. Therefore, ZFN pairs can be
engineered such that DNA binding by each ZFN mediates FokI nuclease dimeri-
zation between ZFP binding sites, resulting in targeted DNA cleavage [38, 39]
(Figure 3.1).
A potential design limitation when designing a ZFP is the lack of availability of
ZF motifs to recognize every triplet sequence [29]. In order to make longer
strings of 6 ZFs, the longer recognition site (18–19 bp) must have ZFs that can
recognize the entire sequence. There are several options available to investiga-
tors for engineering ZFPs for this purpose, and these include modular assembly,
a selection method termed “OPEN,” context-dependent assembly termed
“CoDA,” and a proprietary system available from Sigma-Aldrich. These differing
approaches are discussed in brief later.