Codon swapping of zinc finger nucleases confers expression in primary cells and in vivo from
a single lentiviral vector
Abarrategui-Pontes Cécilia1,2,3, Créneguy Alison1,2,3, Thinard Reynald1,2,3, Fine Eli J.4, Thepenier
Virginie1,2,3, Fournier Laure1,2,3, Cradick Thomas J.4, Bao Gang4, Tesson Laurent
Guillaume5, Anegon Ignacio1,2,3, Nguyen Tuan Huy1,2,3
1 INSERM, UMRS 1064-Center for Research in Transplantation and Immunology,
2 CHU Hôtel Dieu, Institut de Transplantation Urologie Ne´phrologie, Nantes, France
3 Université de Nantes, Faculté de Médecine, Nantes, France
4 Department of Biomedical Engineering, Georgia Institute of Technology and Emory University,
Atlanta, GA 30332, USA
5 Pediatric Hepatogastroenterology - HIFIH laboratory, UPRES 3859, SFR 4038, Angers, France
Corresponding author: Tuan Huy Nguyen, INSERM, UMRS 1064, CHU Hôtel Dieu, 44093 Nantes,
Phone: 33 2 40 08 74 14, Fax: 33 2 40 08 74 11, E-mail: firstname.lastname@example.org
Background: Zinc finger nucleases (ZFNs) are promising tools for genome editing for
biotechnological as well as therapeutic purposes. Delivery remains a major issue impeding targeted
genome modification. Lentiviral vectors are highly efficient for delivering transgenes into cell lines,
primary cells and into organs, such as the liver. However, the reverse transcription of lentiviral vectors
leads to recombination of homologous sequences, such as found between and within ZFN monomers.
Methods: We used a codon swapping strategy to both drastically disrupt sequence identity between
ZFN monomers and to reduce sequence repeats within a monomer sequence. We constructed
lentiviral vectors encoding codon-swapped ZFNs or unmodified ZFNs from a single mRNA transcript.
Cell lines, primary hepatocytes and newborn rats were used to evaluate the efficacy of integrative-
competent (ICLV) and integrative-deficient (IDLV) lentiviral vectors to deliver ZFNs into target cells.
Results: We reduced total identity between ZFN monomers from 90.9% to 61.4% and showed that a
single ICLV allowed efficient expression of functional ZFNs targeting the rat UGT1A1 gene after
codon-swapping, leading to much higher ZFN activity in cell lines (up to 7-fold increase compared to
unmodified ZFNs and 60% activity in C6 cells), as compared to plasmid transfection or a single ICLV
encoding unmodified ZFN monomers. Off-target analysis located several active sites for the 5-finger
UGT1A1-ZFNs. Furthermore, we reported for the first time successful ZFN-induced targeted DNA
double-strand breaks in primary cells (hepatocytes) and in vivo (liver) after delivery of a single IDLV
encoding two ZFNs.
Conclusion: These results demonstrate that a codon-swapping approach allowed a single lentiviral
vector to efficiently express ZFNs and should stimulate the use of this viral platform for ZFN-mediated
genome editing of primary cells, for both ex vivo or in vivo applications.
Key words: Zinc finger nuclease, codon swapping, lentiviral vectors, primary cells, liver, off-target
The ability to modify a genome in a controlled manner is essential for functional genomics, creating
genetically-modified organisms, as well as for industrial and gene therapy purposes. Today, effective
approaches exist to perform controlled genome modifications. These methods are based on the
induction of a site-specific cleavage associated with or without the delivery of an exogenous donor
DNA in order to introduce random or specific mutations, to mark or correct a gene. Development of
genome editing methods relies on the discovery and engineering of artificial endonucleases and
understanding of DNA repair mechanisms [1-6]. Very promising genome editing results were reported
first using meganucleases, then zinc finger nucleases (ZFNs), TALE (transcription activator like
effector) nucleases, and most recently with CRISPR (Clustered Regularly interspersed Short
Palindromic Repeats)/Cas9 [7-10]. ZFNs were created by associating zinc finger domains (ZF) that
target specific DNA sequences and the non-specific cleavage domain from restriction endonuclease
FokI . Each ZF DNA binding domain contains ~30 amino acids and recognizes a specific 3-bp
sequence. Associating 6 ZFs is statistically sufficient to target a unique locus into the human genome.
Because FokI is only active as a dimer, two ZFNs monomers are necessary to perform a specific DNA
double strand break (DSB). The high potency of ZFNs for genome engineering has been
demonstrated in many models: cell lines, primary cells, embryos, and for many species such as rats,
mice, C. elegans, mosquitoes, zebrafish and humans [12-15]. ZFNs hold great promise for treating
genetic diseases, paving the way to personalized medicine [16, 17]. For instance, they allowed precise
gene disruption, conversion, correction and targeted transgene addition in different genomic loci, such
as the AAVS1 safe harbor locus, the IL2-receptor gamma chain (IL2RG) (mutated in X-SCID disease),
and of the alpha1-antitrypsin gene (mutated in alpha-1-antitrypsin deficiency) [18-22]. Nevertheless,
improvements are still needed to increase the effectiveness of ZFN-mediated genome editing,
especially in the context of gene therapy applications. Indeed, a major hurdle is to achieve highly
efficient expression of ZFNs in target cells. Transfection of mRNA or plasmid encoding ZFNs are the
most common delivery methods in cell lines, but these methods are rather inefficient or cytotoxic in
primary quiescent cells and in animals, as compared to viral vectors [23, 24]. In contrast, lentiviral
vectors are very potent tools for delivering transgenes in both quiescent and dividing cells and both in
vitro and in vivo [25-35]. Lentiviral vectors were the first viral platform used for ZFN monomer co-
delivery in target cells using integration-deficient (episomal) lentiviral vectors (IDLV) . In one study,
co-transduction of human cells with two IDLV, each encoding one ZFN monomer of a pair, resulted in
up to 39% of targeted gene editing of the IL2RG gene in the K-562 cell line . However, lentiviral-
mediated targeted gene editing was much less efficient in Epstein-Barr transformed B-lymphocytes,
CD34+ hematopoietic cells and embryonic stem cells . An optimal method to deliver ZFNs into
cells would be to express ZFNs from a single lentiviral vector, as cell co-transduction with two vectors
is rate limiting. Moreover, this would enable stoichiometric expression of both ZFNs monomers.
However, ZFNs monomers are highly homologous (>90%) and it is known that retroviruses recombine
homologous sequences at high frequency during the reverse transcription step . Recently,
Joglekar et al. modified the DNA sequence of one ZFN monomer to decrease its identity to the second
monomer . They reported significant ZFN activity after delivery of the modified pair of ZFNs
expressed from a single IDLV vector into cell lines, but not into primary cells.
In the present work, we used a codon swapping strategy to both drastically disrupt sequence identity
between ZFN monomers and to reduce sequence repeats within a monomer sequence. We showed
that delivery of integrative-competent (ICLV) lentiviral particles encoding both ZFNs in a single vector
resulted in high-level activity of the target UGT1A1 gene in rat cell lines. The high level of activity
allowed characterization of secondary off-target effects of this pair of ZFNs. Furthermore, we reported
for the first time successful ZFN-induced DSB in primary cells (hepatocytes) and in vivo (liver) after
delivery of a single IDLV encoding both ZFNs. Our results should stimulate the use of lentiviral
platform for ZFN-mediated genome editing in primary cells, ex vivo or in vivo.
Materials and Methods
Vector constructs and codon swapping
A pair of custom CompoZr ZFNs targeting the rat UGT1A1 exon 4 (NM_012683.2) was purchased
from Sigma Aldrich (Saint Louis, MO). This pair contains 5 fingers in each monomer and the obligate
heterodimeric ELD/KKR FokI domains . The coding sequence for ZFN Right (ZFN-R) was codon-
swapped according to the following rules: (i) destruction of homologous sequences to ZFN Left (ZFN-
L) that exceed six nucleotides in length, (ii) choice of codons that are mostly expressed in rodent cells
and (iii) removal of consecutive repeat sequences longer than 5 nucleotides. Bicistronic constructs
comprising codon swapped-ZFN-R (swZFN-R) and ZFN-L, or ZFN-R and ZFN-L, separated by a self-
cleaving 2A peptide (T2A) were synthesized by GeneCust (Dudelange, Luxembourg) [39, 40]. These
2-in-1 constructs were subcloned into a self-inactivating HIV1-derived vector
RRLSIN.cPPT.SFFV.GFP.WPRE (kindly provided by Els Verhoyen, INSERM U758, France) in place
of GFP to generate the lentiviral vectors expressing ZFN-L.2A.swZFN-R or ZFN-L.2A.ZFN-R
polyproteins under the control of the spleen focus forming virus (SFFV) promoter, which is highly
active in hepatocytes [41-43]. The packaging plasmid PsPax2-D64V expressing integrase carrying the
mutation D64V was generated by directed mutagenesis of psPAX2 (kindly provided by D Trono,
EPFL, Lausanne, Switzerland). All plasmids were amplified using Nucleobond Xtra Maxi (Macherey
Nagel, Düren, Germany) according to the manufacturer's instructions.
Lentiviral vectors were produced using calcium phosphate-mediated transient transfection of 293T
cells with the vector transfer plasmid, the packaging plasmid (psPAX2, psPAX2 D64V) and the
vesicular stomatitis virus glycoprotein (VSV-G) envelop protein encoding plasmid (pMD2G) as
previously described . Viral supernatants were harvested 24 and 48 hours after transfection and
concentrated by ultracentrifugation at 50,000g for 90 minutes at 4°C in a SW48 rotor. The viral pellet
was resuspended in 100µL of advanced Dubelcco's modified Eagle's medium (Invitrogen, Carlsbad,
CA) and stored at -80°C until use. Vector titers were determined by real-time quantitative PCR on Hela
cells after transduction with serial dilution of virus supernatants. The amplification primers were: GAG-
F 5’-GGAGCTAGAACGATTCGCAGTTA, GAG-R 5’-GGTTGTAGCTGTCCCAGTATTTGTC. For
normalization of the amount of genomic DNA, human beta actin primers were used: HB2-F 5’-
TCCCGTGTGGATCGGCGGCTCCA and H2B-R: 5’-CTGCTTGCTGATCCACATCG. A standard curve
was generated using serial dilutions of a lentiviral vector plasmid in genomic DNA extracted from Hela
cells. Vector titers were calculated as follow: Titer = 1×105 (number of target Hela cells) × (number of
copies per cell of the sample)×(dilution-fold of viral stock for titration)/volume of viral supernatant (in
ml). Vector titers were routinely 1 x109 TU/ml and 5 x 107 TU/ml for ICLV and IDLV vectors,
respectively. Quantification of vector particles, as measured by HIV-1 gag p24 antigen immunocapture
(Retro-TEK kit, Gentaur, Paris, France) showed that concentrated ICLV and IDLV had a similar range
of about 5 x 107 pg of p24/ml.
Cell isolation and culture
The C6 glioma and FAO hepatoma rat cell lines were maintained in Dubelcco's Modified Eagle
Medium (DMEM) (Gibco Life Technologies, Grand Island, NY) supplemented with 10% fetal calf
serum (FCS) (Lonza, Bâle, Switzerland), 2mM glutamine (Sigma Aldrich, St Louis, MO), and
antibiotics (100UI/mL penicillin, 100mg/mL streptomycin, Gibco, Life Technologies, Grand Island, NY).
The 293T cells were maintained in DMEM containing 10% of FCS, 10mM Hepes Gibco (Life
Technologies, Grand Island, NY), 2mM glutamine (Sigma Aldrich, St Louis, MO), and antibiotics
(100UI/mL penicillin, 100mg/mL streptomycin) Gibco (Life Technologies, Grand Island, NY). All cells
were cultured in a humidified atmosphere containing 5% CO2 at 37°C.
Primary rat hepatocytes were isolated from twelve-day-old Wistar pups by a three-step collagenase
perfusion, as described . The liver capsule was then disrupted and cells were mechanically
dispersed in cold DMEM/F-12 Gibco medium (Life Technologies, Grand Island, NY) without serum or
antibiotics. Isolated hepatocytes were purified from non-parenchymal cells by low-speed
sedimentation (50 x g, 5 min) on 30% Percoll solution (GE Healthcare, Little Chalfont, UK). Cells were
resuspended and seeded in DMEM/F-12 Gibco medium containing 10% FBS, 15 mM HEPES Gibco
(Life Technologies, Grand Island, NY), glutamine (Sigma Aldrich, St Louis, MO), antibiotics (100UI/mL
penicillin, 100mg/mL streptomycin) Gibco (Life Technologies, Grand Island, NY). Final cell viability
was greater than 80% as determined by trypan blue exclusion. For primary culture, hepatocytes were
seeded on 24-well collagen IV coated BD BiocoatTM plates (BD Bioscience, Billerica, MA) at a density
of 2 x 105 cells/well. Two hours after seeding, cells were incubated in DMEM/F-12 medium containing
1 x 10-6 M dexamethasone, 1x10-8 M 3,3'-triiodo-L-thyronine, 5 µg/ml bovine insulin (Sigma Aldrich, St
Louis, MO), 5 µg/ml apotransferrin (Sigma Aldrich, St Louis, MO), glutamine (Sigma Aldrich, St Louis,
MO), antibiotics (100UI/mL penicillin, 100mg/mL streptomycin) Gibco (Life Technologies, Grand
Island, NY). The medium was changed daily throughout the culture period.
C6 and FAO cells were transfected at 80-90% confluency using Lipofectamine LTX following
manufacturer’s instructions (Life Technologies, Grand Island, NY). Each well of a 24-well plate was
transfected using 0.5 µg of plasmid DNA. Cells were kept at 37°C for three days before cell analyses.
C6 and FAO cells (5 x104) were seeded in each well of a 6-well plate one day before transduction. On
the day of transduction, the medium was changed before addition of the viral particles. Cells were kept
at 37°C for four days before cell analyses. Primary hepatocytes (2×105) were transduced two hours
after plating by a twelve-hour exposure to lentiviral vector at a multiplicity of infection (MOI) of 60.
Three days later, hepatocytes were harvested with a trypsin-EDTA solution Gibco (Life Technologies,
Grand Island, NY).
Cells were collected for flow cytometry analysis on a LSRII device to evaluate GFP expression using
BDFacsDiva Software for data acquisition (BD Bioscience, Billerica, MA). Data analysis was done with
FlowJo Software (Tree Star, Ashland, OR).
In vivo Vector administration
Wistar rats (Janvier Labs, St Berthevin, France) were maintained on standard laboratory chow and
kept in 12-hour light/dark cycles. All animal experiments were performed according to institutional
guidelines and after approval of the experimental protocols by the Comite d’Ethique des Pays de
Loire. Two-day-old rats were injected with IDLV using a dose of 5 x 106 TU, corresponding to 4 µg of
HIV CAp24 as measured by immunocapture (Retro-TEK kit, Gentaur, Paris, France), in 100 µl of PBS
via the temporal vein.
Genomic DNA was extracted from C6 and FAO cells using the Nucleospin Tissue DNA purification kit
following manufacturer's instructions (Macherey Nagel, Düren, Germany). Genomic DNA from liver
samples was extracted using Blood and Tissue Kit following manufacturer’s instructions (Qiagen,
Valencia, Spain). Primary hepatocytes genomic DNA was extracted with a phenol-chloroform protocol.
Total genomic DNA was quantitated with a spectrophotometer (NanoDrop 2000, Thermo Scientific,
Ottawa, Canada) and quality and contamination were controlled using A260/A280 and A260/A230
T7 endonuclease I (T7EI) mutation detection assay
The percentage of UGT1A1 targeted NHEJ-induced indels (insertion/deletions) was determined using
the T7EI mutation detection assay that measures insertion/deletions and substitutions . Genomic
DNA (100 ng) extracted from cell lines was subjected to PCR with Taq Platinum Hifi (Invitrogen,
Carlsbad, CA) using primers to amplify a 296-bp region of the targeted rat UGT1A1 gene (T7UGT1A1
F 5'- tcacttctctctcccctccc-3' and T7UGT1A1 5'-tttccaggacattcagggtc-3'). The PCR amplification
program was as follows: 94°C for 2 minutes, 40 cycles of 94°C for 30 seconds, 57°C for 30 seconds,
68°C for 30 seconds; and a final 10 minutes at 68°C. Genomic DNA extracted from Liver and primary
hepatocyte was subjected to PCR with AmpliTaq Gold®Fast PCR Master Mix (Applied Biosystems,
Lifetechnologies, Grand Island, NY) using the same primers. These were amplified using: 95°C for 10
minutes, 35 cycles of 96°C for 3 seconds, 57°C for 3 seconds and 68°C for 5 seconds; and a final 10
seconds at 72°C. PCR products were denatured through heating and slow re-annealling allowing
heteroduplex formation using the following conditions: 2 minutes at 95°C, a decrease from 95 to 85°C
(-2°C/second) and a decrease from 85 to 25°C (-0.1°C /second). For cell lines, a final volume of 19µL
of PCR product was digested by 5 units of T7EI (NEB, Beverly, MA, USA) for 30 minutes at 37°C and
separated on 2%-agarose gels. For liver and primary hepatocytes, a final volume of 15µL of PCR
product was digested by 4 units of T7EI (NEB, Beverly, MA, USA) for 30 minutes at 37°C and
separated using capillary gel electrophoresis on Caliper (Perkin Elmer, Waltham, MA) as previously
described 61. T7EI selectively cleaves DNA heteroduplexes, containing re-annealed wild-type and
mutant alleles, or re-annealed alleles with different mutationss due to repair after ZFN cleavage. The
ratio of cleaved (about 118-bp and 114-bp) to uncleaved (296-bp) products was used to calculate
NHEJ frequency as previously described using Image J software . NHEJ frequency was calculated
as follow: %gene modification = 100 X (1-(1-fraction cleaved)^1/2. Labchip GX DNA quantitative
software (Perkin Elmer, Waltham, MA) was used for indels frequency calculation.
Total cell proteins (30 µg/lane) were resolved by electrophoresis and blotted onto nitrocellulose
membranes. After blocking for two hours with TTBS (0.1% Tween, Tween 20-Tris buffered saline)
containing 5% dry milk, the HRP-conjugated murine anti-flag M2 antibodies (1:500, Sigma, Saint
Louis, MO) were added in TBST containing 1% nonfat dry milk. After overnight incubation at 4°C, the
membrane was washed and bands were detected by enhanced chemiluminescence (Pierce
SuperSignal West Pico Chemiluminescent Substrate, Thermo Scientific, Ottawa, Canada). Calnexin
served as loading control and was revealed by using rabbit anti- calnexin (1:000, Enzo Life Sciences,
Farmingdale, NY, USA) and HRP-conjugated anti-rabbit IgG (1:2000, Jackson immunoresearch,
Westgrove, PA, USA) antibodies.
PCR analysis of proviral integrity
Genomic DNA (100 ng) extracted from cell lines was subjected to PCR with Herculase II High Fidelity
Taq (Stratagene, santa Clara, CA) using primers to amplify a 2.8kb region corresponding to full length
ZFN-L.2A-(sw/unmodified)ZFN-R construct (P1-F SFFV: 5’- GGATATCTGCGGTGAGCAG, P4–R
WPRE: CGCTATGTGGATACGCTGC), a 1.5 kb region corresponding to ZFNL (P1-F SFFV, P2-R
T2A 5’-GGATTCTCCTCCACGTCACC-3’) and a 1.3 kb region corresponding to (sw/unmodified)ZFN-
R (P3-F T2A: 5’-CTTCTAACATGCGGTGACGTG-3’, P4-R WPRE). PCR conditions are as follow:
95°C for 2 minutes, 62°C for 5 minutes, 35 cycles of 72°C for 3.5 minutes, 95°C for 30 seconds, 57°C
for 30 seconds; and finally 72°C for 10 minutes.
Off-target effects analysis
The ZFN target site was entered into the PROGNOS online program, which gave output of putative
off-target site lists based on several ranking algorithms . Twenty-one potential off-target sites were
chosen based on a mixture of the top ‘identity’ and ‘Conserved G’s’ PROGNOS rankings algorithms
for ZFNs (Supplementary table 1). Two additional potential off-target sites were chosen based on the
top PROGNOS results for a gapped alignment based on the flexible linker between the 2nd and 3rd
fingers and a ‘rational design’ analysis of the zinc finger binding helices that allowed base degeneracy
at some positions with no binding penalty. PCR and SMRT sequencing analysis of potential off-target
sites was performed exactly as previously described  and using primers listed in supplementary
table 2. Briefly, AccuPrime Taq Hi-Fidelity polymerase (Invitrogen, Carlsbad, CA) was used to amplify
the sites, then samples were purified using MagBind EZ-Pure (Omega Biotek, Norcross, GA), pooled
in roughly equimolar ratios, and sequenced on the PacBio RS using C2/C2 chemistry and consensus
sequencing mode (Pacific Biosciences, Menlo Park, CA). Specific PCR amplification was not achieved
for two of the potential off-target sites and they were discarded from further analysis.
Data are presented as mean ± SD and analyzed using two-tailed Student’s t test (Graph-Pad Prism
Software). P<0.05 was considered significant.
Reduction of Zinc Finger Nuclease identity by codon swapping
As a proof-of-principle, we chose ZFNs targeting the rat UGT1A1 exon 4, which is mutated in
the Gunn rat strain, a spontaneous animal model for Crigler-Najjar disease type I (CNI) . In order
to keep integrity of the ZFN genes after reverse transcription in transduced cells, we used the
redundancy of the genetic code to both change the homologous regions between ZFN-L and ZFN-R
and to swap codons of ZFN-R as described in the materials and methods section. Before codon
swapping, ZFN monomers are highly homologous (>90%) and contain very long homologous
stretches that are up to 333 bp long (Fig. 1A and 2A). By codon swapping of the ZFN-R, we reduced
the total identity to ZFN-L to 61.4% and reduced the length of nearly all identity stretches down to 5 bp
or shorter, with only one 8 bp-long stretch conserved. Identity between ZFN-R and swZFNR was
reduced to 67.7%. We then constructed a single lentiviral vector encoding ZFN-L and swZFN-R
(LV.swZFNs) or the unmodified ZFN pair (LV-ZFNs) as a T2A polyprotein for a stoichiometric
expression of the monomers) [39, 40] (Fig. 1C).
After transfection of C6 rat cells with the plasmids carrying the lentiviral constructs, we
quantified ZFN-driven gene disruption by the T7EI mutation detection assay. This assay reveals
imprecise repair by ZFN-induced non-homologous end joining (NHEJ) leading to substitution
mutations and small insertions and deletions (indels). Transfection efficiency of C6 rat cells was
estimated at 60% using a GFP-encoding plasmid (supplementary Fig. S1). Both LV.ZFNs and
LV.swZFNs plasmids induced significant levels of indels with a higher level for plasmids encoding
unmodified ZFNs (Fig. 2). These data validated the functionality of the lentiviral constructs for targeted
DSB induction in the UGT1A1 locus after transfection. Of note, ZFN activity was undetectable in FAO
cells, probably because of low transfection efficiency as assessed by transfection of GFP-expressing
plasmids (<10% by nucleofection, <5% by lipofection) (data not shown).
Integrity of ZFN-encoding bicistronic vector genomes after lentiviral delivery
We produced ICLV particles carrying the ZFN-encoding bicistronic cassette. Following
transduction of C6 rats cells, we extracted genomic DNA and performed a PCR assay to evaluate the
integrity of proviral genomes using different sets of primers (Fig. 3A). PCR amplicons corresponding to
full-length ZFN-encoding cassettes were detected in cells transduced with LV.swZFNs, while only
PCR amplicons with a deletion of about 1500 bp were detectable in cells transduced with LV.ZFNs
(Fig. 3B). Using sets of primers amplifying the 5’ (P1 and P2 primers) or the 3’ (P3 and P4 primers)
region of the expression cassette, we confirmed the DNA integrity of the proviral LV.swZFNs vector,
and we found that in cells transduced with LV.ZFNs, proviral vectors were re-arranged in the ZFN-R
region (Fig 3B). Plasmids carrying LV.swZFNs and LV.ZFNs yield expected PCR amplicons (not
By Western Blot analyses, we observed that ZFNs migrated as a protein doublet (46.3 and
44.2 kDa) in LV.swZFNs-transduced cells (Fig 3C). This doublet was also reported by others when
ZFNs are expressed as a 2A-polyprotein [50, 51]; it corresponds to the presence of 20 additional
amino acids at the C-terminal end of ZFN-L as compared to ZFN-R. By contrast, in LV-ZFNs
transduced cells, we detected only the upper band of 46 kDa and additional bands of lower molecular
Altogether, the data showed that the lentiviral delivery of a single vector expressing unmodified
ZFNs resulted in sequences rearrangement leading to a deletion in ZFN-R sequence and truncated
ZFNs proteins, while codon swapping allowed maintaining DNA and protein integrities after cell
ZFN activity following lentiviral delivery by a single vector in cell lines
We first determined that almost all C6 and FAO cells (>95%) were transduced after incubation
with GFP-encoding ICLV at a multiplicity of infection (MOI) of 25. We then transduced these cell lines
at a MOI of 25 or a higher with LV.swZFNs or LV.ZFNs and we performed the T7EI mutation detection
assay at day 3 post-transduction. LV-swZFNs were highly efficient at promoting indels events in C6
cells (up to about 30%), as compared to LV-ZFNs (up to about 5%) at a MOI of 25 (Fig. 4A and 4B).
Increasing the MOI allowed the ZFNs to induce indels in up to 60% and 15% of alleles in C6 cells
transduced with LV.swZFNs and LV.ZFNs, respectively. Similarly, we observed that LV.swZFNs
significantly out-performed LV.ZFNs in delivering ZFN activity in FAO cells. In both cell lines, the
difference between LV.swZFNs and LV.ZFNs was more pronounced at a low MOI with a 7-fold
difference in efficiency. Therefore, single lentiviral vectors are highly potent vectors for ZFN delivery
after removing sequence identity using a codon swapping approach.
The high levels of targeted DSBs in the UGT1A1 gene observed after C6 cell transduction with
LV-swZFNs (up to 60%, Fig. 4b) gave the opportunity to look for potential off-target effects, which may
not be easily detectable with a low ZFN activity, as observed after transfection. Off-target sites were
chosen for investigation based on a mixture of the top-ranked sites by the PROGNOS method 
(Supplementray Table 1). Using SMRT sequencing, we interrogated 22 genomic loci that were highly
ranked for cleavage in the rat genome by the Ugt1A1 ZFNs (the intended target site and 21 potential
"off-target" sites) in transduced C6 cells using pairs of primers listed in supplementary table 2 .
High levels of modification (about 64.4%) at the UGT1A1 intended target was confirmed (Table 1). We
also detected off-target activity at 6 additional loci (Supplementary Fig. S1 and Supplementary Table 3
for sequence reads of NHEJ). Off-target sites were found in an intronic region of the Tarsl2 gene at a
frequency of about 10% and in five other sites at low frequency. Of note, Tarsl2 and Gbas off-target
sites share only 80% sequence identity with the target site.
ZFN-mediated editing in primary cells and in vivo by single IDLV and LV
Based on the high efficiency observed in cell lines, we treated primary rat hepatocytes with integrative
competent and integrative-deficient lentiviral particles at a MOI of 60. Of note, hepatocytes were
cultured in non-stimulating growth conditions, in absence of serum and growth factors. We detected
significant levels of targeted indels in the UGT1A1 gene after cell transduction with IDLV-swZFNs
(about 2%) (Fig. 5 and Supplementary Fig. 2). A higher ZFNs activity was observed after transduction
with their integrative counterparts (5-15%). We then investigated whether IDLV-swZFNs were able to
impart ZFN activity in vivo. For this purpose, we administrated lentiviral particles to rat neonates. At
day 10 post-injection, we performed the T7EI mutation detection assay on liver samples. About 1% of
alleles contained indels indicative of targeted DSBs in 4 out of 5 injected animals (Fig. 6 and
Supplementary Fig. 3).
In this study, we developed a strategy to allow the delivery of ZFNs from a single lentiviral
vector. Such vectors are able to efficiently transduce a wide range of cells or organs, and in particular
mitotically quiescent cells. The use of integration-deficient lentiviral vectors allows transient transgene
expression in dividing cells, and prolonged transgene expression in slowly or non-dividing cells, such
as retina, adult liver and muscle cells [30, 52-56]. Despite their potency, few studies reported the use
of the lentiviral platform for delivering ZFNs. ICLV and IDLV were shown to drive co-delivery of ZFNs
monomers into cell lines and primary cells [19, 37, 57, 58]. For nuclease mediated gene editing, the
preferable scenario is transient targeted DSB activity in transduced cells to minimize off-target effects
and cytotoxicity. Although the simplest way to transfer a ZFNs pair might be to co-transduce target
cells with two separate vectors, it is difficult to ensure the co-transduction of the same cells,
particularly when transduction efficiency is low, and pre-selection is not possible. In addition, in vivo
applications, such as neonatal vector delivery, are limited by the small volume of vector preparation
that can be injected. Conceivably, use of a single construct to express ZFNs pair might enhance the
efficiency of genome editing. For instance, in several studies, ZFNs were expressed from a single
plasmid, adenoviral or baculoviral construct [51, 59-62]. For these reasons and because of a low
transgene expression from IDLV and of an expected low in vivo hepatic transduction efficacy, we
aimed to construct a 2-in-1 lentiviral vector to ensure co-expression of ZFNs monomers in transduced
cells. Nonetheless, in highly permissive cell lines and with high-titer vector preparations, a 2-in-1 ZFNs
construct might not be more efficient than two single ZFN constructs .
Because ZFN monomers are highly homologous, it is anticipated that retroviral vectors are not
adequate to express both monomers from a single mRNA transcript. Indeed, Holkers et al. showed
that a TALE nuclease monomer, which contains highly repetitive regions, is rearranged during reverse
transcription leading to mutated TALE nucleases in lentivirally-transduced cells . Using codon-
optimization to minimize internal repeated sequences, it was possible to deliver TALE nuclease
monomer by a lentiviral vector . Recently, a codon-optimization approach was used to reduce
identity between ZFN-L and ZFN-R from 95% to 85%. This allowed inducing successful ZFN-targeted
DSBs from a single IDLV in cell lines, but not in primary cells such as T lymphocytes or CD34+ cells
despite the use a strong promoter driving ZFN expression [37, 65]. Those authors proposed that this
failure in primary cells may be overcome by further increasing transgene expression from IDLV.
Notably, it has been shown that IDLV are subjected to an epigenetic gene silencing and that treating
transduced cells with histone deacetylase (HDAC) inhibitors, such as trichostatin A, significantly
stimulated transgene expression in different cell types, including primary cells and regardless of the
proliferation status .
In the present study, we hypothesize that optimal lentiviral delivery of ZFNs expressed from a
single transcript requires minimal identity between ZFNs monomers as well as a minimal length of
homologous stretches. Using a codon swapping approach, we reduced total ZFNs identity to 61.4%
and reduced identity stretches to 5 bp at maximum, except one 8-bp stretch remaining in the FokI
domain. Previous studies showed that lentiviral vectors can tolerate short homologous repeats [28, 33,
66]. We confirmed sequence rearrangements of ZFNs after cell transduction with LV encoding
unmodified ZFNs. These recombination events occurred in the ZFN-R region leading to proteins with
reduced molecular weights. Of note, we detected a residual activity (about 15% at a MOI of 200)
suggesting that sequence rearrangement led to the deletion of one or two repeats but a functional 3-
or 4- finger ZFN-R was still delivered. By contrast, we observed structural integrity of ZFNs at a DNA
and protein level in cells transduced with LV.swZFNs. Unexpectedly, expression of swZFNs was lower
than that of ZFNs at a protein level, even though amino-acid sequence of ZFN-L should be identical in
both constructs (Fig. 3C). It would be interesting to elucidate how the rearranged ZFN-R can increase
overall ZFNs protein stability or expression, but it is out of scope of this study. LV.swZFNs out-
performed LV.ZFNs by up to 7-fold at equivalent MOI with up to about 60% of NHEJ-induced indels at
the targeted site in C6 cells and FAO cells. Although we transduced almost all the C6 (Supplementary
Fig. 1) and FAO cells (data not shown), we observed that FAO cells are much less sensitive to gene
modification. As previously described, indel levels and sensitivity to ZFNs differ between cell types [38,
67-69]. This may be due to differences in transduction, transgene expression efficiencies, and/or in
homology-directed recombination activity. Altogether, our data confirmed that high identity between
ZFN-L and ZFN-R is a major hurdle when using lentiviral vectors as a delivery platform. Our results
are in agreement with the study of Joglekar et al., in which recombination between triple flag tag and
nuclear localization signal sequences was detected . This could also have occurred in our study,
as LV.ZFNs constructs contain triple flag tags. As compared to plasmid transfection, our results
highlight the potency of lentiviral transduction, which induced 15% of indels events in C6 cells. This
reflects a high transduction rate (almost 100% efficiency) leading to a high level of transgene
expression as compared to a lower transfection efficiency (60%) and transgene expression (2-log
difference) (Supplementary Fig. S1). Furthermore, plasmid transfection was inefficient in FAO cells
(data not shown) and failed to induce detectable levels of indels in contrast to lentiviral transduction.
When nucleases cut at other locations in the genome other than their intended target, they can
potentially induce unwanted gene disruption, destabilization of the cell’s genome, or transformation of
the cell into a cancerous phenotype. Therefore, it is important to reveal potential off-target sites of a
given pair of ZFNs. A pre-requisite for detecting off-target sites is to induce sufficient on-target activity.
Because of high ZFN activity in C6 cells, we could investigate off-target effects using the PROGNOS
algorithm in these cells . Given that only 21 potential off-target sites were interrogated, finding six
bona fide locations of off-target activity, validated as having a statistically significant mutation
frequency greater than untreated cells, was higher than expected. Based on off-target studies of other
3- and 4-finger ZFNs, an analysis of this size would be expected to yield perhaps four bona fide off-
target sites [48, 70-73]. Moreover, this study is the first report of bona fide off-target activity discovered
for 5-finger ZFNs with heterodimeric FokI domains. In our study, the levels of on-target activity after
lentiviral transduction were much higher than in previous studies . Thus, it is probable that
conducting these studies by delivering ZFNs using a single lentiviral vector and our codon swapping
approach revealed off-target sites that wouldn’t have been detected otherwise with an approach that
yielded lower on-target activity. Our data highlight the impact of the nuclease design and of the
efficacy of delivery methods on off-target effects and thus on the biosafety of artificial nucleases for
gene therapy purposes.
The liver is a privileged target organ for gene therapy. For such a purpose, IDLV are the preferable
ZFN-delivery platform as compared to ICLV, as IDLV allow transient expression of ZFNs. After
lentiviral neonatal delivery, ZFN expression is expected to be short because of the dilution of IDLV
genomes in the highly dividing hepatocytes, as observed after neonatal delivery of AAV vectors [50,
74]. In adult mice livers, transgene expression from IDLV slowly decreased overtime, probably
because of slow turn-over of hepatocytes or IDLV silencing [30, 58]. It is noteworthy that no toxicity
was observed after delivery of AAV vectors encoding ZFNs into the liver of adult mice . In primary
rat hepatocytes, cultured using serum-free and growth-factor free conditions, we observed indels at
levels of about 2.5% and 5-15% after transduction with IDLV and ICLV, respectively. The lower indel
frequency induced by IDLV is consistent with a lower transgene expression from IDLV as compared to
ICLV [30, 76, 77]. It has been reported that lowering the temperature to 30°C after ZFNs delivery
increased indel frequency, and that TSA or sodium butyrate improved transgene expression from
IDLV [38, 58]. However, we observed that in ICLV-transduced primary rat hepatocytes cultured at
30°C or treated with sodium butyrate, there was no increase in indel frequency and that TSA (10 µM)
was toxic (data not shown). Stimulated by in vitro results with IDLV, we evaluated the in vivo
performance of IDLV-swZFNs in the liver of rat neonates. We previously showed that after neonatal
delivery, the liver is the first target organ and that most of transduced cells in the liver are hepatocytes
. In all but one injected animals, we detected indels events at a level of about 1%, which is close to
that observed in primary rat hepatocytes. Because ZFNs-induced toxicity has been associated with
cleavage at off-target sites [61, 79, 80], this suggests that in vivo genotoxicity would be very low, if
any. To our knowledge, our data represent the first description of effective ZFN-mediated gene editing
using a single IDLV vector as a delivery platform in primary cells and in vivo in the liver. Furthermore,
we also demonstrated targeted DSB of an endogenous natural genomic locus. Previous studies of in
vivo gene editing in the liver using ZFNs or meganucleases were performed with transgenic animals in
which the nuclease targeted sequence was knocked into the mouse genome and were thus artificial
[50, 75, 81]. These artificial targeting genomic loci may differ in chromatin status for nuclease
accessibility and/or gene transcription as compared to a targeted endogenous locus. Notably, we
observed a similar level of targeted DSBs using IDLV.swZFNs to that reported after AAV-mediated
delivery of ZFNs carrying the obligate heterodimeric ELD/KKR FokI (as we used in the present study)
into the liver . This low level of ZFN-targeted DSBs was sufficient to cure hemophilic mice, in which
an AAV vector carrying a donor template for correcting a mutated human factor IX mini-gene was co-
injected with AAV encoding ZFNs . Other applications for which a low level of gene editing could
achieve therapeutic effects are conceivable. For instance, we previously showed that as few as 0.03
vector copy/haploid genome of retroviral vectors encoding UGT1A1 were sufficient to normalize the
bilirubinemia of the Gunn rat, which is the spontaneous model of Crigler-Najjar disease type I . The
target site for the ZFNs used in this study is located in the rat UGT1A1 exon 4, close to the genetic
mutation responsible for the disease in this animal model. In a previous study of in vivo ZFN-mediated
genome editing in hemophilic mice, homology-directed repair events accounted for about 20% of the
total gene modification events. Thus, the level of in vivo ZFN-mediated targeted DSBs reported here
could be sufficient to correct, at least partially, the metabolic defect of the Gunn rat by co-delivering
ZFNs and a DNA donor template for gene repair or for targeted insertion of a therapeutic expression
cassette using the HDR pathway. Complete metabolite correction would require stimulation of the
HDR pathway, such as by treating Gunn rats with vanillin, a potent inhibitor of NHEJ, that increased in
vivo hepatic gene targeting frequencies . Liver diseases in which there is a selective growth
advantage of corrected hepatocytes may be privileged diseases, such as hereditary tyrosine type I.
Indeed, as few as 0.01-0.04% corrected hepatocytes rescued the phenotype of mice modeling
hereditary tyrosine type I [83, 84]. Finally, lentiviral delivery of ZFNs alone can be useful to study, for
instance, the effect on in vivo invalidation of potential tumor suppressor genes in tumor initiation and
growth as their inactivation will confer a selective growth advantage, or for the in vitro invalidation of
CCR5 or CXCR4 in CD4+ T cells to confer resistance to HIV-1 infection [61, 85].
In conclusion, our study describes a codon-swapping strategy that allowed efficient delivery of
ZFNs by a single lentiviral vector and highlights the potency of this viral delivery platform. The
improved activity was likely due to minimizing recombinogenic events occurring during reverse
transcription by the reduced ZFN identity and elimination of long stretches of homologous sequences.
We also used the SFFV promoter to drive ZFNs expression, which may be stronger than the ones
used in the study of Joglekar et al. (intronless human elongation factor 1a promoter, Moloney murine
Leukemia LTR with myeloproliferative sarcoma virus enhancer) . The non-proliferative status of
our target cells, i.e. primary hepatocytes, in which ZFNs (vector genome and proteins) are not diluted
by cell divisions, may have contributed though there were also significant indels in neonate
proliferating livers. The codon-swapping strategy is broadly applicable to ZFNs, as the monomer
sequences are usually very similar except for the codons corresponding to positions -1 to 6 of the
DNA binding helices. Either integrative competent and non-integrative lentiviral vectors can be used,
depending on the application. Because of the potency of lentiviral vectors, we anticipate that high-
levels of ZFN activity could be achieved in various cell lines that are refractory to transfection (such as
FAO cells), primary cells or organs in vivo. Because ICLV are much more potent than IDLV, as we
observed in this study and also reported by others in the liver , doxycycline-regulated ICLV or ICLV
encoding destabilized ZFNs could be advantageously used [86-88]. Future directions include
demonstrating therapeutic efficacy of ZFN-mediated in vivo genome editing using a single IDLV as a
viral delivery platform, for instance in the Gunn rat. The present study should stimulate the use of
lentiviral vectors for delivering ZFNs, as a potential viral platform to modify cell genomes and to treat a
wide range of genetic disease, by either ex vivo or in vivo gene therapy approaches.
Abbreviations: AAV, adeno associated virus; CRISPR, Clustered Regularly interspersed Short
Palindromic Repeats; DMEM, Dubelcco's Modified Eagle Medium, DSB, double strand break
FCS fetal calf serum; HDAC, histone desacetylase; HDR, homology-directed repair; ICLV integrative-
competent lentiviral vectors; IDLV, integrative-deficient lentiviral vectors; LTR, long terminal repeats;
MOI, multiplicity of infection; NHEJ, non homologous end joining; SFFV, spleen focus forming virus;
Sw, swap; TALE, nuclease transcription activator like effector nucleases; TSA, trichostatin A;
UGT1A1, UDP-glucoronosyl-transferase A1; ZF, zinc finger domains; ZFN-L, zinc finger nuclease left;
ZFN-R, zinc finger nuclease right; ZFN, zinc finger nuclease.
Funding supports: This work was supported by the Association Française contre les Myopathies
(AFM), the Association Francophone des Glycogénoses), the Association Française pour l’étude du
foie (AFEF), the ‘‘Agence Nationale de la Recherche” and IHU-Cesti, which received French
government financial support managed by the National Research Agency via the "Investment Into The
Future" program ANR-10-IBHU-005. The IHU-Cesti project is also supported by Nantes Metropole and
the Région Pays de la Loire. This work was also supported by the National Institutes of Health as an
NIH Nanomedicine Development Center Award [PN2EY018244 to GB]. E.J.F. was in addition
supported by the National Science Foundation Graduate Research Fellowship [DGE-1148903].
Disclosure of interest: The authors declare no conflict of interest.
Author contributions: AC, CA, TR, FE, TV, FL, and TL performed the experiments; AC, FE, CT, BG,
PG, AI and NTH analyzed data and drafted the paper; AI, PG and NTH supervised study and wrote
chromosomes by using the I-SceI system of Saccharomyces cerevisiae. Mol. Cell. Biol 1995;15:1968-
2. Donoho, G, Jasin M and Berg P. Analysis of gene targeting and intrachromosomal homologous
recombination stimulated by genomic double-strand breaks in mouse embryonic stem cells. Mol. Cell.
3. Jasin, M. Genetic manipulation of genomes with rare-cutting endonucleases. Trends Genet.
4. Jasin, M and Berg P. Homologous integration in mammalian cells without target gene
selection. Genes Dev. 1988;2:1353-1363.
5. Porteus, MH and Baltimore D. Chimeric Nucleases Stimulate Gene Targeting in Human Cells.
6. Rouet, P, Smih F and Jasin M. Expression of a site-specific endonuclease stimulates
homologous recombination in mammalian cells. Proc. Natl. Acad. Sci. U.S.A 1994;91:6064-6068.
7. Concorder, JP and Giovannangeli C. Engineered nucleases for targeted genome modification.
Curr Gene Ther 2011;11:1.
8. Gaj, T, Gersbach CA and Barbas CF, 3rd. ZFN, TALEN, and CRISPR/Cas-based methods for
genome engineering. Trends Biotechnol. 2013.
9. Mashimo, T. Gene targeting technologies in rats: Zinc finger nucleases, transcription activator-
like effector nucleases, and clustered regularly interspaced short palindromic repeats. Dev. Growth
10. Perez-Pinera, P, Ousterout DG and Gersbach CA. Advances in targeted genome editing. Curr
Opin Chem Biol 2012;16:268-277.
11. Kim, YG, Cha J and Chandrasegaran S. Hybrid restriction enzymes: zinc finger fusions to Fok I
cleavage domain. Proc. Natl. Acad. Sci. U.S.A 1996;93:1156-1160.
12. Carroll, D. Zinc-finger nucleases: a panoramic view. Current gene therapy 2011;11:2-10.
13. Geurts, AM, Cost GJ, Freyvert Y, et al. Knockout rats via embryo microinjection of zinc-finger
nucleases. Science 2009;325:433.
14. Remy, S, Tesson L, Menoret S, et al. Zinc-finger nucleases: a powerful tool for genetic
engineering of animals. Transgenic research 2009;19:363-371.
15. Tesson, L, Usal C, Menoret S, et al. Knockout rats generated by embryo microinjection of
TALENs. Nat Biotechnol 2011;29:695-696.
16. Carroll, D. Progress and prospects: zinc-finger nucleases as gene therapy agents. Gene Ther.
17. Rahman, SH, Maeder ML, Joung JK, et al. Zinc-finger nucleases for somatic gene therapy: the
next frontier. Human gene therapy 2011;22:925-933.
18. Hockemeyer, D, Soldner F, Beard C, et al. Efficient targeting of expressed and silent genes in
human ESCs and iPSCs using zinc-finger nucleases. Nat Biotechnol 2009;27:851-857.
19. Lombardo, A, Genovese P, Beausejour CM, et al. Gene editing in human stem cells using zinc
finger nucleases and integrase-defective lentiviral vector delivery. Nat Biotechnol 2007;25:1298-1306.
20. Tay, FC, Tan WK, Goh S-L, et al. Targeted transgene insertion into the AAVS1 locus driven by
baculoviral vector-mediated zinc finger nuclease expression in human-induced pluripotent stem cells.
The journal of gene medicine 2013;15:384-395.
21. Urnov, FD, Miller JC, Lee YL, et al. Highly efficient endogenous human gene correction using
designed zinc-finger nucleases. Nature 2005;435:646-651.
22. Yusa, K, Rashid ST, Strick-Marchand H, et al. Targeted gene correction of alpha1-antitrypsin
deficiency in induced pluripotent stem cells. Nature 2011;478:391-394.
23. Chuah, MK and Vandendriessche T. Optimizing delivery and expression of designer nucleases
for genome engineering. Human gene therapy methods 2013;24:329-332.
24. Schott, JW, Galla M, Godinho T, et al. Viral and non-viral approaches for transient delivery of
mRNA and proteins. Curr Gene Ther 2011;11:382-398.
Choulika, A, Perrin A, Dujon B, et al. Induction of homologous recombination in mammalian
patients with Wiskott-Aldrich syndrome. Science 2013;341:1233151.
26. Annoni, A, Cantore A, Della Valle P, et al. Liver gene therapy by lentiviral vectors reverses anti-
factor IX pre-existing immunity in haemophilic mice. EMBO molecular medicine 2013;5:1684-1697.
27. Biffi, A, Montini E, Lorioli L, et al. Lentiviral hematopoietic stem cell gene therapy benefits
metachromatic leukodystrophy. Science 2013;341:1233158.
28. Brown, BD, Venneri MA, Zingale A, et al. Endogenous microRNA regulation suppresses
transgene expression in hematopoietic lineages and enables stable gene transfer. Nat Med
29. Genovese, P, Schiroli G, Escobar G, et al. Targeted genome editing in human repopulating
haematopoietic stem cells. Nature 2014;510:235-240.
30. Matrai, J, Cantore A, Bartholomae CC, et al. Hepatocyte-targeted expression by integrase-
defective lentiviral vectors induces antigen-specific tolerance in mice with low genotoxic risk.
31. Naldini, L, Blomer U, Gallay P, et al. In vivo gene delivery and stable transduction of
nondividing cells by a lentiviral vector. Science 1996;272:263-267.
32. Provasi, E, Genovese P, Lombardo A, et al. Editing T cell specificity towards leukemia by zinc
finger nucleases and lentiviral gene transfer. Nat Med 2012;18:807-815.
33. Schmitt, F, Remy S, Dariel A, et al. Lentiviral vectors that express UGT1A1 in liver and contain
miR-142 target sequences normalize hyperbilirubinemia in Gunn rats. Gastroenterology
34. Zhou, Y, Zhu S, Cai C, et al. High-throughput screening of a CRISPR/Cas9 library for
functional genomics in human cells. Nature 2014;509:487-491.
35. Zufferey, R, Donello JE, Trono D, et al. Woodchuck hepatitis virus posttranscriptional
regulatory element enhances expression of transgenes delivered by retroviral vectors. J Virol
36. Weiss, R, Teich N, Varmus HE, et al. RNA Tumor Viruses. Cold Spring Harbor Laboratory.
New York, 1984 1984.
37. Joglekar, AV, Hollis RP, Kuftinec G, et al. Integrase-defective lentiviral vectors as a delivery
platform for targeted modification of adenosine deaminase locus. Molecular therapy: the journal of the
American Society of Gene Therapy 2013;21:1705-1717.
38. Doyon, Y, Vo TD, Mendel MC, et al. Enhancing zinc-finger-nuclease activity with improved
obligate heterodimeric architectures. Nature methods 2011;8:74-79.
39. Szymczak, AL and Vignali DAA. Development of 2A peptide-based strategies in the design of
multicistronic vectors. Expert Opin Biol Ther 2005;5:627-638.
40. Yang, S, Cohen CJ, Peng PD, et al. Development of optimal bicistronic lentiviral vectors
facilitates high-level TCR gene expression and robust tumor cell recognition. Gene Ther
41. Themis, M, Waddington SN, Schmidt M, et al. Oncogenesis following delivery of a nonprimate
lentiviral gene therapy vector to fetal and neonatal mice. Mol Ther 2005;12:763-771.
42. Waddington, SN, Nivsarkar MS, Mistry AR, et al. Permanent phenotypic correction of
hemophilia B in immunocompetent mice by prenatal gene therapy. Blood 2004;104:2714-2721.
43. Weber, K, Bartsch U, Stocking C, et al. A multicolor panel of novel lentiviral "gene ontology"
(LeGO) vectors for functional gene analysis. Mol Ther 2008;16:698-706.
44. Bovia, F, Salmon P, Matthes T, et al. Efficient transduction of primary human B lymphocytes
and nondividing myeloma B cells with HIV-1-derived lentiviral vectors. Blood 2003;101:1727-1733.
45. Nguyen, TH, Birraux J, Wildhaber B, et al. Ex vivo lentivirus transduction and immediate
transplantation of uncultured hepatocytes for treating hyperbilirubinemic Gunn rat. Transplantation
46. Mussolino, C, Morbitzer R, Lutge F, et al. A novel TALE nuclease scaffold enables high
genome editing activity in combination with low toxicity. Nucleic Acids Res 2011;39:9283-9293.
47. Guschin, DY, Waite AJ, Katibah GE, et al. A rapid and general assay for monitoring
endogenous gene modification. Methods in molecular biology (Clifton, N.J 2010;649:247-256.
Aiuti, A, Biasco L, Scaramuzza S, et al. Lentiviral hematopoietic stem cell gene therapy in
TALE nuclease off-target cleavage. Nucleic acids research 2013.
49. Ikushiro, S. Takashi Iyanagi: UGT1 gene complex: from Gunn rat to human. Drug Metab. Rev
50. Li, H, Haurigot V, Doyon Y, et al. In vivo genome editing restores haemostasis in a mouse
model of haemophilia. Nature 2011;475:217-221.
51. Miller, JC, Holmes MC, Wang J, et al. An improved zinc-finger nuclease architecture for highly
specific genome editing. Nat Biotechnol 2007;25:778-785.
52. Apolonia, L, Waddington SN, Fernandes C, et al. Stable Gene Transfer to Muscle Using Non-
integrating Lentiviral Vectors. Mol Ther 2007;15:1947-1954.
53. Hutson, TH, Verhaagen J, Yáñez-Muñoz RJ, et al. Corticospinal tract transduction: a
comparison of seven adeno-associated viral vector serotypes and a non-integrating lentiviral vector.
Gene Ther. 2012;19:49-60.
54. Loewen, N, Leske DA, Chen Y, et al. Comparison of wild-type and class I integrase mutant-FIV
vectors in retina demonstrates sustained expression of integrated transgenes in retinal pigment
epithelium. J Gene Med 2003;5:1009-1017.
55. Philippe, S, Sarkis C, Barkats M, et al. Lentiviral vectors with a defective integrase allow
efficient and sustained transgene expression in vitro and in vivo. Proc Natl Acad Sci U S A
56. Rahim, AA, Wong AMS, Howe SJ, et al. Efficient gene delivery to the adult and fetal CNS using
pseudotyped non-integrating lentiviral vectors. Gene Ther. 2009;16:509-520.
57. Lombardo, A, Cesana D, Genovese P, et al. Site-specific integration and tailoring of cassette
design for sustainable gene transfer. Nature methods 2011.
58. Pelascini, LP, Maggio I, Liu J, et al. Histone Deacetylase Inhibition Rescues Gene Knockout
Levels Achieved with Integrase-defective Lentiviral Vectors Encoding Zinc-finger Nucleases. Human
gene therapy methods 2013.
59. Duda, K, Lonowski LA, Kofoed-Nielsen M, et al. High-efficiency genome editing via 2A-coupled
co-expression of fluorescent proteins and zinc finger nucleases or CRISPR/Cas9 nickase pairs.
Nucleic Acids Res 2014;42:e84.
60. Lei, Y, Lee CL, Joo KI, et al. Gene editing of human embryonic stem cells via an engineered
baculoviral vector carrying zinc-finger nucleases. Mol Ther 2011;19:942-950.
61. Perez, EE, Wang J, Miller JC, et al. Establishment of HIV-1 resistance in CD4+ T cells by
genome editing using zinc-finger nucleases. Nat Biotechnol 2008;26:808-816.
62. Straimer, J, Lee MC, Lee AH, et al. Site-specific genome editing in Plasmodium falciparum
using engineered zinc-finger nucleases. Nature methods 2012;9:993-998.
63. Holkers, M, Maggio I, Liu J, et al. Differential integrity of TALE nuclease genes following
adenoviral and lentiviral vector gene transfer into human cells. Nucleic acids research 2013;41.
64. Yang, L, Guell M, Byrne S, et al. Optimization of scarless human stem cell genome editing.
Nucleic Acids Res 2013;41:9049-9061.
65. Joglekar, AV, Stein L, Ho M, et al. Dissecting the Mechanism of Histone Deacetylase Inhibitors
to Enhance the Activity of Zinc Finger Nucleases Delivered by Integrase-Defective Lentiviral Vectors.
Hum Gene Ther 2014.
66. Brown, BD, Gentner B, Cantore A, et al. Endogenous microRNA can be broadly exploited to
regulate transgene expression according to tissue, lineage and differentiation state. Nat Biotechnol
67. Connelly, JP, Barker JC, Pruett-Miller S, et al. Gene Correction by Homologous Recombination
With Zinc Finger Nucleases in Primary Cells From a Mouse Model of a Generic Recessive Genetic
Disease. Molecular Therapy 2010;18:1103-1110.
68. Herrmann, F, Garriga-Canut M, Baumstark R, et al. p53 Gene repair with zinc finger nucleases
optimised by yeast 1-hybrid and validated by Solexa sequencing. PloS one 2011;6.
69. Rousseau, J, Chapdelaine P, Boisvert S, et al. Endonucleases: tools to correct the dystrophin
gene. The Journal of Gene Medicine 2011;13:522-537.
Fine, EJ, Cradick TJ, Zhao CL, et al. An online bioinformatics tool predicts zinc finger and
nuclease specificity. Nature Biotechnology 2011.
71. Gupta, A, Meng X, Zhu LJ, et al. Zinc finger protein-dependent and -independent contributions
to the in vivo off-target activity of zinc finger nucleases. Nucleic Acids Research 2011;39:381-392.
72. Pattanayak, V, Ramirez CL, Joung JK, et al. Revealing off-target cleavage specificities of zinc-
finger nucleases by in vitro selection. Nature methods 2011;8:765-770.
73. Sander, JD, Ramirez CL, Linder SJ, et al. In silico abstraction of zinc finger nuclease cleavage
profiles reveals an expanded landscape of off-target sites. Nucleic acids research 2013.
74. Flageul, M, Aubert D, Pichard V, et al. Transient expression of genes delivered to newborn rat
liver using recombinant adeno-associated virus 2/8 vectors. J Gene Med 2009;11:689-696.
75. Anguela, XM, Sharma R, Doyon Y, et al. Robust ZFN-mediated genome editing in adult
hemophilic mice. Blood 2013;122:3283-3287.
76. Bayer, M, Kantor B, Cockrell A, et al. A large U3 deletion causes increased in vivo expression
from a nonintegrating lentiviral vector. Mol Ther 2008;16:1968-1976.
77. Okada, Y, Ueshin Y, Hasuwa H, et al. Targeted gene modification in mouse ES cells using
integrase-defective lentiviral vectors. Genesis (New York, N.Y.: 2000) 2009;47:217-223.
78. Nguyen, TH, Aubert D, Bellodi-Privato M, et al. Critical assessment of lifelong phenotype
correction in hyperbilirubinemic Gunn rats after retroviral mediated gene transfer. Gene Ther
79. Mussolino, C and Cathomen T. On target? Tracing zinc-finger-nuclease specificity. Nature
80. Radecke, S, Radecke F, Cathomen T, et al. Zinc-finger nuclease-induced gene repair with
oligodeoxynucleotides: wanted and unwanted target locus modifications. Mol Ther 2010;18:743-753.
81. Gouble, A, Smith J, Bruneau S, et al. Efficient in toto targeted recombination in mouse liver by
meganuclease-induced double-strand break. J Gene Med 2006;8:616-622.
82. Paulk, NK, Loza LM, Finegold MJ, et al. AAV-mediated gene targeting is significantly enhanced
by transient inhibition of nonhomologous end joining or the proteasome in vivo. Hum Gene Ther
83. Paulk, NK, Wursthorn K, Wang Z, et al. Adeno-associated virus gene repair corrects a mouse
model of hereditary tyrosinemia in vivo. Hepatology 2010;51:1200-1208.
84. Yin, H, Xue W, Chen S, et al. Genome editing with Cas9 in adult mice corrects a disease
mutation and phenotype. Nat Biotechnol 2014.
85. Yuan, J, Wang J, Crain K, et al. Zinc-finger nuclease editing of human cxcr4 promotes HIV-1
CD4(+) T cell resistance and enrichment. Mol Ther 2012;20:849-859.
86. Lachmann, N, Brennig S, Pfaff N, et al. Efficient in vivo regulation of cytidine deaminase
expression in the haematopoietic system using a doxycycline-inducible lentiviral vector system. Gene
87. Pruett-Miller, SM, Reading DW, Porter SN, et al. Attenuation of zinc finger nuclease toxicity by
small-molecule regulation of protein levels. PLoS Genet 2009;5:e1000376.
88. Szulc, J, Wiznerowicz M, Sauvain MO, et al. A versatile tool for conditional gene expression
and knockdown. Nature methods 2006;3:109-116.
Gabriel, R, Lombardo A, Arens A, et al. An unbiased genome-wide analysis of zinc-finger
Table 1. SMRT Sequencing confirms on- and off-target activity at sites ranked by PROGNOS
The 22 top-ranked cleavage sites of UGT1A1 ZFNs, as determined by the PROGNOS 'identity' and
'Conserved G's' algorithms, were evaluated using SMRT sequencing in transduced C6 cells. The
'match type' indicates the orientation of the left (L) and right (R) nucleases at the site and the length of
the spacer sequence. Site sequences are listed as 5'—(+) half-site—spacer—(-) half-site—3'.
Therefore, the (+) positive half-site is listed in the reverse anti-sense orientation as compared to the
DNA sequence that the ZFN binds. In sequences, lower-case letters indicate mismatches as
compared to the target site. “C6 cell line modification frequency” is the percentage of observed
sequences showing evidence of NHEJ events. For all sites shown, significantly higher (p < 0.05)
frequencies of indels were observed in transduced cells as compared to mock treated cells. Our
analysis revealed that the locus near the Ftcd gene contains a homozygous C->T SNP at the second
position of the (-) half-site, introducing an additional mutation in that ZFN binding site relative to the
Figure 1. Codon swapping of ZFNs targeting rat UGT1A1 and vector design. (A) Schematic of
sequence alignment of ZFN-L to ZFN-R (left) and ZFN-L to swZFN-R (right). ZFN-L sequence is
represented by horizontal black bar and, ZFN-R and swZFN-R by grey bar; vertical lines represent
identical nucleotides and nucleotides number is depicted. (B) The percentage of sequence identity
between ZFN-L and ZFN-R, and ZFN-L and swZFN-R are indicated. (C) Schematic of the single
lentiviral vector plasmid constructs. “U3-R-U5”: HIV-1–derived self-inactivating long terminal repeat, ψ:
HIV-1 packaging signal, SFFV: spleen focus forming virus promoter, T2A: 2A peptide from Thosea
assigna, WPRE: woodchuck hepatitis virus post-transcriptional regulatory element, “ΔU3-R-U5”: HIV-
1–derived self-inactivating long terminal repeat containing U3 in which viral enhancer and promoter
have been deleted (SIN vector).??
Figure 2. Functionality of ZFN-encoding lentiviral plasmids. (A) Rat C6 cells were transfected
using plasmids carrying single lentiviral constructs encoding unmodified ZFNs (LV.ZFNs) or codon-
swapped ZFNs (LV.swZFNs) using Lipofectamine LTX. Three days after transfection, genomic DNA
was isolated and subjected to T7EI mutation detection assay in the targeted UGT1A locus site. T7E1
DNA products were migrated. Asterisk and arrows indicate uncleaved (296 bp) and cleaved products
(about 182 bp and 114 bp), respectively. The frequency of gene modification was then calculated
according to the ratio of cleaved to uncleaved DNA products, as described in the Materials and
Methods. Numbers below each lane indicate the percentage of targeted indels. NT: non-transfected
cells, L: 1Kb DNA ladder. (B) The frequency of targeted allelic disruption was evaluated at 3-days
post-transfection and calculated from 3 independent experiments. Bars indicate mean ± SEM.
Figure 3. Structural analyses of ZFNs in transduced C6 cells. (A) Primers (P1 to P4) for testing
integrity of ZFNs-expression cassette are depicted. (B) PCR analyses of transduced cells. Cells were
transduced with integrative competent LV.swZFNs or LV.ZFNs vector at an MOI of 200. Three days
after transduction, total DNA was extracted and PCRs were performed with P1-P4 (top), P1-P2
(middle) and P3-P4 (bottom) primers pair. Arrows indicate the expected length of PCR amplicons. NT:
non-transduced cells. L: 1 kb DNA ladder. (C) ZFN protein expression in transduced cells. Total
proteins were extracted three days after transduction. Western blot against the FLAG tag revealing
ZFN proteins (top) and calnexin (90 kDa) as a loading control protein (bottom) are shown. Asterisks
indicate the expected 46 kDa and 42 kDa ZFN protein bands and # indicate the unexpected ZFN
protein band (37 kDa).
Figure 4. Activity of ZFNs delivered by a single lentiviral vector into cell lines. (A) In vitro on-
target cleavage in transduced cells. C6 cells were transduced with LV.swZFNs (swZFNs) or LV.ZFNs
(ZFNs) at the indicated MOI. Three days post-transduction, total DNA was isolated and subjected to
T7 endonuclease I assay for detecting UGT1A1-targeted DSBs. Migration of cleaved (arrows) and
uncleaved (asterisk) products is indicated. Numbers below each lane indicate the percentage of
targeted indels. NT: non-transduced cells, L: 1Kb DNA. (B) The frequency of targeted allelic disruption
of UGT1A1 gene was evaluated at 3-days post-transduction in C6 (left) and FAO (right) cells.
Experiments have been repeated in duplicate three times. Bars indicate mean ± SEM. n.d.: not
Figure 5. Lentiviral delivery of ZFNs into primary hepatocytes. Primary hepatocytes from five rats
(R1 to R5) were isolated and transduced with IDLV-swZFNs or integrative competent LV-swZFNs at
an MOI of 60. Three days post-transduction, total DNA was isolated and subjected to T7EI mutation
detection assay. T7EI DNA products were separated using capillary gel electrophoresis. The
frequency of targeted UGT1A1 allelic disruption was shown for hepatocytes isolated from different rats
and transduced with IDLV (R1-IDLV, R2-IDLV, R3-IDLV) and ICLV (R4-LV, R5-LV); Bars indicate
mean ± SEM.
Figure 6. Delivery of ZFN by a single IDLV into the liver. Two-day-old rats were injected with IDLV-
swZFNs (5 x 106 TU). Ten days after injection, genomic DNA was extracted and subjected to T7EI
mutation detection assay. T7E1 DNA products were separated using capillary gel electrophoresis. The
frequency of targeted UGT1A1 allelic disruption was shown for individual newborn rats (n=5, NB1 to
Page 38 Download full-text