THE JOURNAL OF BIOLOGICAL CHEMISTRY
© 2005 by The American Society for Biochemistry and Molecular Biology, Inc.
Vol. 280, No. 5, Issue of February 4, pp. 3707–3714, 2005
Printed in U.S.A.
Exploring Strategies for the Design of Artificial
Transcription Factors
TARGETING SITES PROXIMAL TO KNOWN REGULATORY REGIONS FOR THE INDUCTION OF ␥-GLOBIN
EXPRESSION AND THE TREATMENT OF SICKLE CELL DISEASE*
Received for publication, June 17, 2004, and in revised form, October 28, 2004
Published, JBC Papers in Press, November 10, 2004, DOI 10.1074/jbc.M406809200
Torbjörn Gräslund‡, Xuelin Li‡, Laurent Magnenat‡, Mikhail Popkov, and Carlos F. Barbas III§
From the Skaggs Institute for Chemical Biology and the Department of Molecular Biology, The Scripps Research Institute,
La Jolla, California 92037
Artificial transcription factors can be engineered to
interact with specific DNA sequences to modulate endogenous gene expression within cells. A significant
hurdle to implementation of this approach is the selection of the appropriate DNA sequence for targeting. We
reasoned that a good target site should be located in
chromatin, where it is accessible to DNA-binding proteins, and it should be in the close vicinity of known
transcriptional regulators of the gene. Here we have
explored the efficacy of these criteria to guide our selection of potential regulators of ␥-globin expression. Several zinc finger-based transcriptional activators were
designed to target the sites proximal to the ⴚ117-position of the ␥-globin promoter. This region is proximal to
the binding sites of known and potential natural transcription factors. Design and study of three transcription factors identified the potent transcriptional activator, gg1-VP64-HA. This transcription factor was able to
interact directly with the ␥-globin promoter and upregulate expression of reporter gene constructs as well
as the endogenous gene in a selective manner. Transfection of a gg1-VP64-HA expression vector or retroviral
delivery of this transcription factor into the erythroleukemia cell line K562 resulted in an increase of fetal
hemoglobin. The ␥-globin content of cells expressing
gg1-vp64-HA showed up to 16-fold higher levels of fetal
hemoglobin than the native K562 cell line. These transcriptional activators constitute a novel class of regulators of the globin locus that may be suitable for treatment of diseases arising from mutations in this locus
such as sickle cell disease and thalassemic diseases.
The human ␤-globin locus is located on chromosome 11 and
contains five homologous globin genes that are differentially
expressed during development (Fig. 1). In the embryonic stage
of development, the ⑀-gene is predominantly expressed. Its
expression is replaced with ␥-globin during the fetal stage;
␥-globin forms tetrameric fetal hemoglobin (HbF),1 consisting
* This work was supported by National Institutes of Health Grant 1
RO1 DK61803-01 and by a postdoctoral fellowship from the WennerGren foundations. The costs of publication of this article were defrayed
in part by the payment of page charges. This article must therefore be
hereby marked “advertisement” in accordance with 18 U.S.C. Section
1734 solely to indicate this fact.
‡ These authors contributed equally to this work.
§ To whom correspondence should be addressed: Dept. of Molecular
Biology, Skaggs Institute for Chemical Biology, BCC-550, The Scripps
Research Institute, 10550 N. Torrey Pines Rd., La Jolla, CA 92037. Tel.:
858-784-9098; Fax: 858-784-2583; E-mail: carlos@scripps.edu.
1
The abbreviations used are: HbF, fetal hemoglobin; DBD, DNAThis paper is available on line at http://www.jbc.org
of two ␥- and two ␣-chains. Shortly after birth, the expression
of ␥-globin starts to decline, and it is gradually replaced by
␤-globin, which forms ␣2␤2, the adult version of hemoglobin
(HbA). Transcriptional control of these genes is mediated by
silencers and enhancers both within and flanking the genes in
the locus (1–5). Situated upstream of the genes is a locus
control region (LCR) that is necessary for activation of the
entire locus. The differential expression of the globin genes is
thought to depend on physical competition of the genes for
association with the LCR. This association is mediated by
different transcription factors (6, 7). Several genetic diseases
are associated with defective ␤-globin expression, the most
common being ␤-thalassemias and sickle cell disease.
A feature common to both sickle cell disease and ␤-thalassemia
is that disease can be mitigated by the production of the fetal
globin chain ␥ that can serve to form ␣2␥2 or fetal hemoglobin
HbF. One example of such protection from disease is the benign
condition known as “hereditary persistence of fetal hemoglobin”
(HPFH), which is characterized by an increase in the level of fetal
hemoglobin. Fetal hemoglobin is normally 1–2% of the globin
level, but in HPFH it can be increased up to 20%. Patients that
have both SCD and HPFH typically have minimal symptoms of
disease (8, 9). These observations have spurred studies to induce
HbF expression for the treatment of SCD, and several potent
compounds have been identified including hydroxyurea, butyrate, ribavirin, activin A, and 5-azacytidine (10). Treatment with
these compounds, however, produces a range of side effects. Another possible therapeutic intervention for patients suffering
from SCD or thalassemia is to deliver globin genes directly (11,
12) or to impart continuous transcriptional activation of the endogenous HbF gene through gene therapy approaches. In contrast to treatment with blood transfusions or with compounds
like hydroxyurea, these approaches would potentially allow for
treatment of the diseases only once or rarely.
In this study, we have investigated the use of artificial transcription factors as a means of inducing ␥-globin expression.
Phage-based selection and rational design have shown that Cys2His2 zinc fingers can be adapted to bind a wide variety of DNA
sequences (13–18). Typically, a zinc finger domain recognizes
three base pairs of DNA, and recognition of all possible DNA
triplets therefore requires an alphabet of 64 different domains.
Many such specific DNA-binding domains are now available (16 –
binding domain; MBP, maltose-binding protein; TF, transcription factor; LCR, locus control region; HPFH, hereditary persistence of fetal
hemoglobin; SCD, sickle cell disease; HA, hemagglutinin; ChIP, chromatin immunoprecipitation; PBS, phosphate-buffered saline; DAPI,
4⬘,6-diamidino-2-phenylindole; KRAB, Krüppel-associated box; ICAM,
intercellular adhesion molecule; EGFP, enhanced green fluorescent
protein.
3707
3708
Targeted Up-regulation of Human ␥-Globin
18). In contrast to most transcription factors that rely on dimerization of protein domains for extending protein-DNA contacts to
longer DNA sequences, or addresses, simple covalent linkage of
tandem repeats of the zinc finger domain allow for the recognition of longer asymmetric sequences of DNA. Three to six domains are typically fused to yield a DNA-binding domain (DBD)
recognizing 9 –18 bases of DNA. These polydactyl zinc finger
proteins typically target DNA sequences with high affinity and
specificity. When fused to activation or repression domains, these
proteins become potent regulators of the transcriptional activity
of the target gene. One activation domain that has shown promise as a potent activator when fused to zinc finger protein is the
VP64 domain. The VP64 domain consists of a tetrameric repeat
of the minimal activation domain (amino acids 437– 447) of the
herpes simplex protein VP16 (19).
Artificial transcription factors have been applied to regulate
endogenous gene expression in a number of instances (20 –25).
A significant obstacle to creating productive artificial transcription factors is the selection of the DNA site to be targeted.
We reasoned that a good target site should be located in chromatin, where it is accessible to DNA-binding proteins and it
should be in the close vicinity of known transcriptional regulators of the gene. There have been reports correlating the
ability of designed three-finger proteins in regulating gene
expression to the location of their target sites in accessible
chromatin structures (26, 27). Accessible target sites may be
located in the linker region between positioned nucleosomes
and are sensitive to micrococcal nuclease digestion or are hypersensitive to DNase I. In the promoter region of the human
␥-globin genes, no DNase I-hypersensitive sites have been
found; nor have the positions of nucleosomes been determined.
However, the promoter sequences required for the ␥-globin
expression have been studied extensively. These sequences
may represent the binding sites for transcriptional regulators
and thus are probably located in accessible chromatin regions.
Therefore, we chose target sequences next to known cis-regulatory sites with our zinc finger proteins. Target sites immediately adjacent to the known cis-regulatory sites may be advantageous in gene regulation as well. It has become clear that in
a gene regulation event, both transcriptional activation functions and repression functions are imposed on the promoter,
and the regulatory outcome depends on the relative strengths
of these opposing functions (28). Closely positioned activators
on the promoter often act synergistically to boost transcription.
Therefore, by following these guidelines, it may be possible to
effectively select target sites that are both accessible and supportive of synergism within activators.
Here we have studied the potential of these design selection
criteria for the development of artificial transcription factors for
the targeted up-regulation of ␥-globin expression. We have analyzed the ability of our designed transcription factors to activate
transcription of reporter constructs and the endogenous ␥-globin
gene in the human erythroleukemia cell line K562.
MATERIALS AND METHODS
Assembly of the DNA-binding Domains and Construction of Expression Vectors for the Transcription Factors—Recognition helices for fingers recognizing 5⬘-GNN-3⬘ or 5⬘-ANN-3⬘ triplets of DNA have been reported (16 –18). Assembly of DBDs, each consisting of six zinc finger
domains fused head-to-tail, were performed as previously described (17,
29, 30) and were subcloned into the bacterial expression vector pMALCSS, a derivative of the bacterial expression vector pMAL-c2 (New England Biolabs, Beverly, MA). The DBDs were subsequently subcloned into
pcDNA3-VP64-HA, as described elsewhere (29), using the two SfiI restriction sites. The resulting plasmids each express a fusion protein with an
N-terminal DBD followed by the transcriptional activator domain VP64
and a C-terminal hemagglutinin (HA) decapeptide tag.
Expression, Purification, and DNA Binding Characterization of the
DBD—The DBD were expressed as maltose-binding protein (MBP)
fusion proteins from the pMAL-CSS constructs. The proteins could be
purified to ⬎90% purity using the Protein and Fusion purification
system (New England Biolabs) as per the manufacturer’s protocol,
except that the running buffer in the column purification was zinc
buffer A (10 mM Tris-HCl, pH 7.5, 90 mM KCl, 1 mM MgCl2, 90 ␮M
ZnCl2) supplemented with 1% bovine serum albumin and 5 mM dithiothreitol. Protein purity was analyzed by separating eluted proteins on a
15% SDS-polyacrylamide gel. Protein concentrations were determined
using the Bio-Rad protein assay and was compared with a bovine serum
albumin standard. The dissociation constant for the DBD and its target
sequence was determined using an electrophoretic mobility shift assay
(EMSA) essentially as described (31).
Construction of a Dual Luciferase Reporter Plasmid—The ␥␤-2luc
plasmid was constructed by fusing the ␥- and ␤-globin promoters to the
firefly and Renilla luciferase genes, respectively. Others have reported
a similar reporter vector (32). Firefly luciferase was isolated from the
pGL3-basic vector (Promega, Madison, WI) by digestion with HindIII
and SalI. It was inserted into pRL-null (Promega) that had been cut
with the same enzymes, resulting in the plasmid p-2luc. The ␤-globin
promoter from position ⫺206 to ⫹47 relative to the start of transcription was PCR-amplified using the primers ␤-3b (5⬘-GAACTGCAGTCTGTTTGAGGTTGCTAGTGAACACAG-3⬘) and ␤-5f (5⬘-GGAATTCAAGAGATATATCTTAGAGGGAGGGCTG-3⬘). The resulting PCR product
and the vector p-2luc were cut with EcoRI and PstI and ligated to give
the vector ␤R-2luc. The ␥-globin promoter from position ⫺206 to ⫹47
relative to the start of transcription was PCR-amplified using the primers ␥-3b (5⬘-CCCAAGCTTGGACTAGGAGCTTATTGATAATCTCAG3⬘) and ␥-5f (5⬘-TCACCGCGGGCCCCTTCCCCACACTATCTCAATG3⬘) and was cut with SacII and HindIII. The DNase I-hypersensitive
sites 2 and 3 (LCRHS3 and LCRHS2) comprise bases 8489 – 8865 and
4425– 4823 of the human ␤-globin locus. LCRHS2 was PCR-amplified
using LCRHS2–3b (5⬘-TCCCCGCGGCTAGAATATGTCACATTCTGTCTCAGG-3⬘) and LCRHS2–5f (5⬘-GAGGGTACCGCTTCAGTTTTTCCTTAGTTCCTGTTAC-3⬘), and was cut using KpnI and SacII. LCRHS3
was PCR-amplified using LCRHS3–3b (5⬘-CGGGGTACCCTGGGGGTATAGGGGAGCAGTCCCATG-3⬘) and LCRHS3–5f (5⬘-CCGCTCGAGCTGTGGATAAAGGAGACAGCCATGAAG-3⬘) and was cut using XhoI
and KpnI. The vector ␤R-2luc was cut using XhoI and HindIII. Cut
fragments corresponding to ␤R-2luc, LCRHS2, LCRHS3, and ␥-globin
promoter were ligated, resulting in the vector ␥␤-2Luc.
Cell Lines—All cell culture experiments were performed using the
erythroleukemia cell line K562 (ATCC number CCL-243). Cells were
maintained at a density of ⬍4 ⫻ 105/ml in RPMI 1640 medium (Invitrogen) supplemented with 10% fetal calf serum (Hyclone, Logan, UT) and
1% antibiotic antimycotic mixture (Invitrogen). Transfection of DNA
constructs to the cells was mediated by PLUS Reagent and
LipofectamineTM (Invitrogen).
Analysis of the Transcription Factors (TFs) Using a Reporter Assay—
For each sample, different amounts of the pcDNA-DBD-VP64-HA construct (2–50 ng) was mixed with 50 ng of the reporter plasmid ␥␤-2luc
and 50 ng of the CMV-LacZ reporter plasmid. The amount of plasmid in
each sample was adjusted to a total of 150 ng by the addition of the
starting pcDNA3.1 plasmid. Each sample was then transfected into
50 ⫻ 103 K562 cells in a 96-well plate format. As a control, hemin
(Sigma) was used in different concentrations to treat K562 cells that
had been transfected with 50 ng of ␥␤-2luc and 50 ng of CMV-LacZ.
Extracts from cells transfected with TFs were prepared ⬃72 h posttransfection, and extracts from cells treated with hemin were prepared
⬃96 h post-transfection. Growth medium was changed on hemin
treated cells 48 and 72 h after transfection. The cell lysates were
analyzed for their content of firefly and Renilla luciferase by the DualLuciferase reporter assay system (Promega). The cell lysates were also
analyzed for their content of ␤-galactosidase using the Galacto-Light
Plus™ reporter system (Tropix, Bedford, MA). The luciferase values
could subsequently be normalized to the ␤-galactosidase activity.
Theoretical Analysis of gg1 Specificity—The programs used for the
theoretical analysis of the ␤-globin locus were BLASTN with the option,
“Search for short, nearly exact matches” (available on the World Wide
Web at www.ncbi.nlm.nih.gov/BLAST/) and Fuzznuc, written by Alan
Bleasby (available on the World Wide Web at www.hgmp.mrc.ac.
uk/Software/EMBOSS/Apps/fuzznuc.html).
Regulation of Endogenous ␥-Globin Expression—K562 cells (1.2 ⫻
107) were co-transfected with 12 ␮g of pcDNA-gg1-VP64-HA or 12 ␮g of
pcDNA3.1 (Invitrogen) and 3 ␮g of pLEGFP (BD Biosciences) and
grown for 72 h with one exchange of growth medium after 48 h. The
cells were subsequently sorted in a FACSvantage DiVa (BD Biosciences), and 10% of the cells with the highest enhanced green fluorescent protein signal were collected and analyzed by Western blotting.
Targeted Up-regulation of Human ␥-Globin
Retroviral expression of gg1 transcription factors was accomplished by
subcloning gg1 into pMX-effector vectors. Virus was prepared and delivered as described (20, 25). Selection of cell lines with stable expression of gg1-VP64-HA was performed by growing cells transfected with
pcDNA-gg1-VP64-HA for 5 days, after which Geneticin G418 (Invitrogen) was added to a concentration of 250 ␮g/ml. Cells were maintained
for an additional 14 days, after which individual surviving cells were
allowed to expand in the presence of 250 ␮g/ml Geneticin, and the
resulting clonal populations were analyzed.
Analysis of HbF Content by Western Blotting—Samples were collected by centrifugation and resuspended in reporter lysis buffer (Promega, Madison, WI). Cells were incubated on ice for 20 min followed by
freezing at ⫺80 °C. Cells were thawed on ice, and solid material was
removed by centrifugation. Cell lysate containing 10 ␮g of protein was
separated on a 15% SDS-polyacrylamide gel followed by blotting to a
nitrocellulose filter. The filter was probed with a mouse monoclonal
anti-HbF antibody (Caltag Laboratories, Burlingame, CA) at 0.5 ␮g/ml
in Tris-buffered saline supplemented with 0.1% Tween 20. The filters
were developed using the ECL⫹plus kit (Amersham Biosciences).
DNase I Footprinting—For DNase I footprinting, 100 pmol of the
oligonucleotide fp-1 (5⬘-TGTCTGAAACGGTCCCTGGCTAA-3⬘) was
end-labeled with 32P by incubation with 1 ␮l [␥-32P]ATP (8000 Ci/mmol)
and 10 units of T4 polynucleotide kinase (Roche Applied Science) in T4
polynucleotide kinase buffer for 60 min at 37 °C. Fp-1 was subsequently
used with the oligonucleotide fp-2 (5⬘-ACTGCTGAAGGGTGCTTCCTT3⬘) in a PCR. The PCR mixture was separated on a nondenaturing
polyacrylamide gel, and the band corresponding to the 180-bp fragment
encompassing the gg1-binding site was recovered. For each footprinting
reaction, the following was mixed: ⬃1 ng of 32P-labeled PCR product
and from 100 nM to 1 pM MBP-gg1 in 390 ␮l of zinc buffer A supplemented with 0.5 mM dithiothreitol. The reactions were equilibrated
overnight at 4 °C and then digested with DNase I. The digested samples
were separated on a 6% polyacrylamide gel containing 8.3 M urea, 88
mM Tris borate (pH 8.3), and 2 mM EDTA. The gel was developed using
a phosphor imager screen and subsequently visualized in a Molecular
Dynamics SF phosphor imager.
Chromatin Immunoprecipitation (ChIP)—ChIP assays were performed as described with the following modifications (25, 33). Nuclei
from a minimum of 107 cells/immunoprecipitation were cross-linked,
isolated, and sonicated to produce 500-bp chromatin fragments.
TFZF-chromatin complexes were captured by centrifugation using 10 ␮l
of ZF antibody or no antibody and lyophilized StaphA cells (Calbiochem), washed extensively and analyzed by PCR for 35 cycles using the
forward 5⬘-GACTGAATCGGAACAAGGCAAAG-3⬘ and reverse 5⬘TGTCCTCCTCTGTGAAATGACCC-3⬘ primers producing a 338-bp PCR
fragment containing the gg1 target site in the human ␥-globin gene
locus. As negative control, PCR primers amplifying the ICAM-1 promoter were used as described (25).
Analysis of HbF Content by Flow Cytometry—Samples of ⬃5 ⫻ 105
cells were collected and washed with 100 ␮l PBS supplemented with
0.1% bovine serum albumin (PBS-B). Cells were fixed by incubation for
10 min in 0.05% glutaraldehyde (Sigma) in PBS-B followed by washing
with PBS-B. After fixation, cells were permeabilized by incubation with
0.1% Triton X-100 (Sigma) in PBS-B for 3 min followed by washing with
PBS-B. Samples were stained by incubation with R-phycoerythrinconjugated mouse monoclonal antibody against fetal hemoglobin (PEHbF) (Caltag Laboratories) for 45 min at a working concentration of 2.5
␮g/ml in PBS-B followed by washing with PBS-B. Cells stained for HbF
content were analyzed by flow cytometry using a FACScalibur instrument (BD Biosciences).
Analysis of mRNA Levels of gg1-VP64 Using RT-PCR and Real Time
PCR—Samples containing 106 cells were collected followed by RNA
extraction using the TriReagent kit (Molecular Research Center, Cincinnati, OH). After the last precipitation step, the pellet containing pure
RNA was resuspended in 150 ␮l of water and frozen at ⫺80 °C until
analyzed. For RT-PCR analysis of gg1-VP64-HA, 2.5 ␮g of RNA was
annealed with an oligo(dT) primer and then reverse transcribed with
the Superscript kit (Invitrogen). For each RT-PCR analysis, 5% of the
synthesized cDNA was used as a template in a nested PCR. The first
PCR was with the outer primers TF-1 (5⬘-ACTGCCGCGACCTTGCT-3⬘)
and TF-2 (5⬘-CGGAACGTCGTACGGGTAGTT-3⬘) using the following
program: denaturation (94 °C, 30 s) followed by 25 cycles of amplification (94 °C, 30 s; 55 °C, 30 s; 72 °C, 60 s) and ending with a final
extension step (72 °C, 7 min). 1 ␮l of the outer PCR was used as a
template for the inner PCR with the primers TF-3 (5⬘-TTCTCCCGCAGCGATCAC-3⬘) and TF-4 (5⬘-CCAAAGCACCTGGGTCTGA-3⬘). The
PCR program for the inner PCR was the same as for the outer, except
that the 72 °C extension step in each amplification cycle was for 30 s
3709
instead of 60 s. RT-PCR of GADPH was performed using gadph-f (5⬘CCATGTTCGTCATGGGTGTGA-3⬘) and gadph-r (5⬘-CATGGACTGTGGTCATGAGT-3⬘). The PCR program used was the same as the inner
PCR of gg1-VP64-HA.
The amount of ␥-globin was quantified by a real time PCR analysis of
the synthesized cDNA. The analysis was carried out on an ABI PRISM
7700 Sequence Detection System (Applied Biosystems, Foster City, CA)
according to the manufacturer’s protocols, using SYBR Green dye for
detection of synthesized amplicons. The primers used for ␥-globin
were ggrt-f (5⬘-GACAAGCTGCATGTGGATCCT-3⬘) and ggrt-r (5⬘CCGAAATGGATTGCCAAAAC-3⬘). This primer pair gave an amplicon
that extended over the splice junction between exon 2 and exon 3. The
primers used for GADPH were gadph-f and gadph-r. All values for the
␥-globin mRNA were normalized against the amount of GADPH mRNA.
Immunocytochemical Cellular Localization of gg1-VP64-HA and
HbF—Cells were seeded on poly-L-lysine-coated Lab-Tek cover glasses
and incubated for 30 min. Cells were washed with copious amounts of
PBS, incubated with 4% paraformaldehyde for 20 min, washed again,
and incubated in a humidifying chamber at room temperature for 1 h
with primary antibody mixture of biotinylated rat anti-HA monoclonal
antibody (5 ␮g/ml) (Roche Applied Science) and mouse anti-human HbF
monoclonal antibody (Caltag Laboratories). The cells were then stained
for 1 h at room temperature with the mixture of Cy5-conjugated donkey
anti-mouse IgG polyclonal antibodies and streptavidin/rhodamine
Red-X (both from Jackson Immunoresearch, West Grove, PA) diluted to
1:100 in PBS-B, 0.1% saponin. Finally, the cells were incubated with
DAPI solution for 5 min, washed with PBS, and covered with SlowFade
antifade reagent. Three-color (DAPI, rhodamine Red-X, and Cy5) threedimensional data sets were collected with a DeltaVision system (Applied Precision, Issaquah, WA); this consisted of an Olympus IX-70
fluorescence microscope, a motorized high precision xyz stage, a 100watt mercury lamp, and a KAF1400 chip-based cooled charge-coupled
device camera. Exposure times were 0.2– 0.5 s (2-binning), and images
were obtained with a ⫻60 oil objective. Three-dimensional reconstruction was generated by capturing 150-nm serial sections along the z axis.
Images were deconvolved (based on the Agard-Sadat inverse matrix
algorithm) and analyzed with softWorX version 2.5.
RESULTS
Targeting the ␥-Globin Promoter with Designed DNA-binding Domains—We designed and assembled the genes encoding
three hexameric zinc finger-based DBDs to interact with the
A
␥- and G␥-globin promoters (henceforth referred to as the
␥-globin promoter). These DBDs, called gg1, gg2, and gg3, were
designed to interact with the ␥-promoter in a region where
other cis-acting TFs bind. Their predicted sites of interaction
with the ␥-promoter and the sites of binding of the endogenous
␥-globin transcriptional activators FKLF and FKLF-2 are
shown in Fig. 1A. After assembly, the DBDs were expressed in
Escherichia coli as a fusion to MBP, and crude extracts from
the cell cultures were analyzed for DNA binding in a sandwich
enzyme-linked immunosorbent assay experiment. All three
were found to bind strongly and specifically to the 18-base pair
target DNA sequences for which they were designed (data not
shown). The MBP-DBD fusion proteins were readily purified to
homogeneity on amylose resin. The dissociation constants for
the purified MBP-DBD fusion proteins and their target sequences were determined in an electrophoretic mobility shift
assay, and the results are listed in Table I. The dissociation
constant for gg1 and its target sequence was 0.7 nM. In contrast, the dissociation constants for gg2 and gg3 were 5.3 and
5.4 nM, respectively, almost 10 times higher than that for gg1.
Analysis of the TFs Using a Reporter Assay—Transcription
factors were prepared from the three zinc finger proteins
through fusion with the VP64 activation domain and were
expressed using the expression vector pcDNA-VP64-HA. The
resulting plasmids expressed transcription factors consisting of
an N-terminal zinc finger protein fused to the domain VP64
and a C-terminal HA decapeptide tag that facilitates tracking
of the expressed protein (Fig. 1B). In order to analyze the
ability of the designed TFs to activate transcription of the ␥globin gene, a reporter vector ␥␤-2luc was constructed. The
Targeted Up-regulation of Human ␥-Globin
3710
FIG. 1. A (top), the 80-kb ␤-globin locus
located on chromosome 11. A (bottom), a
detailed map of the A␥-promoter from position ⫺145 to ⫺87 relative to the start of
transcription. The binding sites of gg1
and gg2 (binding the bottom strand) as
well as gg3 (binding the top strand) are
shown as arrows. B, schematic view of
how the artificial transcription factors are
composed. The DBD is at the N terminus
followed by VP64, and the HA tag is located at the C terminus.
TABLE I
Characteristics of the constructed DBDs
DBD
Target sequence
关Kd兴a
gg1
gg2
gg3
5⬘-GTC AAG GCA AGG CTG GCC-3⬘
5⬘-AGG CTA TTG GTC AAG GCA-3⬘
5⬘-GAC AAG GCA AAC TTG ACC-3⬘
0.7
5.3
5.4
a
The dissociation constant of the DBD from its target sequence
expressed in nM.
␤-globin promoter was incorporated into the reporter vector in
order to assess the specificity of gene activation. The ␥␤-2luc
vector is shown in Fig. 2A. The vector encodes the ␥-globin
promoter fragment sequence from ⫺206 to ⫹47 relative to the
start of transcription and drives expression of the firefly luciferase gene. The ␤-globin promoter fragment consists of the
sequence from ⫺206 to ⫹47 relative to the start of transcription and drives expression of the Renilla luciferase gene. Also
included in ␥␤-2luc were the DNase I-hypersensitive sites 2
and 3 from the ␤-globin LCR. This portion of the LCR contains
enhancer elements important for high level activation of the
entire ␤-globin locus (34).
Vector DNAs expressing the transcriptional activators were
co-transfected with ␥␤-2luc into K562 cells. The amount of
expressed firefly and Renilla luciferase was measured 72 h post
transfection, and the results are plotted in Fig. 2, B and C, as
a function of the amount of transfected expression vector. Included in the experiment was a control that expressed e2cVP64-HA (29). This transcriptional activator was not predicted
to interact with either the ␥- or the ␤-promoters specifically. As
a control, K562 cells transfected with ␥␤-2luc were stimulated
with hemin at different concentrations. There was a dose-dependent response for both the ␥-globin and ␤-globin promoters
(data not shown).
gg1-VP64-HA was the strongest transcriptional activator, and
luciferase expression was increased 47-fold over background at
the highest concentration of transfected plasmid (Table II). By
contrast, gg2-VP64-HA and gg3-VP64-HA failed to activate luciferase expression of the reporter construct even at the highest
plasmid concentrations. The negative control activated luciferase
expression slightly. Hemin could activate this reporter construct
moderately and had a 22-fold up-regulation of luciferase expression at the highest concentration relative to untreated cells. To
examine the specificity of gg1-VP64-HA in this reporter assay, we
created a reporter construct containing two mutated sites within
the designed target site of gg1, A9G and G12T, but otherwise
identical to ␥␤-2luc. These mutations were designed to compromise interactions of the third and the fourth fingers of the polydactyl zinc finger protein with their target sites. Indeed, these
mutations destroyed the ability of gg1-VP64-HA to activate the
reporter gene, whereas a reporter construct displaying only the
G12T mutation was activated at an intermediate level (data not
shown).
To analyze whether the TFs could interact with the ␤-globin
promoter and activate ␤-globin expression, we also measured
the levels of Renilla luciferase in the transfected cells. gg1VP64, gg2-VP64-HA, and gg3-VP64-HA induced Renilla luciferase expression 4-, 1-, and 1-fold relative to untreated cells at
the highest plasmid concentrations. This was similar to the
activation observed with negative controls. Hemin activated
luciferase expression 2-fold.
DNase I Footprinting of gg1—Since gg1-VP64-HA was the
only TF that could activate expression by interaction with the
␥-globin promoter in the reporter assay, it was further analyzed. To ascertain that gg1 bound to the predicted target
sequence within the ␥-globin promoter, a DNase I footprinting
experiment was performed with a stepwise dilution series of
MBP-gg1 from 100 nM to 1 pM; the resulting gel is shown in Fig.
3. This study demonstrated that gg1 bound to the 18-bp site
targeted within the ␥-globin promoter.
Theoretical Analysis of gg1 Specificity—The LCR and proximal promoters (from ⫺2000 relative to the start of transcription up to and including the untranslated region) of all genes in
the ␤-globin locus were analyzed to identify putative gg1 interaction sites using BLASTN and Fuzznuc. A BLASTN search
did not identify any sites within the LCR or promoters with
more than 8-base homology to the predicted gg1 binding site,
except for the two targeted sites within the A␥ and G␥ promoter.
Several sites with 8-base or lower homology were found evenly
dispersed throughout the LCR and promoters. Fuzznuc, which
allows any base mismatch within the search sequence, did not
Targeted Up-regulation of Human ␥-Globin
3711
TABLE II
Up-regulation by the transcriptional activators in a reporter assay
Data below are based on values from Fig. 2, B and C, at 50 ng of
transfected plasmid or 50 ␮M hemin. The values are an average of three
independent measurements.
DBD
Activation of ␥-globin
Activation of ␤-globin
-fold
gg1
gg2
gg3
e2c
Hemin
47 ⫾ 5
1 ⫾ 0.2
1 ⫾ 0.1
7⫾2
22 ⫾ 7
4 ⫾ 0.5
1 ⫾ 0.1
1 ⫾ 0.1
8⫾2
2 ⫾ 0.1
FIG. 3. DNase I footprinting of the ␥-globin promoter at different concentrations of gg1. The numbers to the right of the gel
indicate the position of DNase I cleavage relative to the start of transcription. The position of the gg1 binding site is also shown to the right.
The numbers above each lane correspond to the concentration of MBPgg1 in the sample. Lane C, a negative control, where no protein was
added. Lane L, a ladder.
FIG. 2. A, schematic map of the reporter plasmid ␥␤-2luc. This plasmid was used to evaluate the activating potential of the artificial
transcription factors. It contains the ␥-globin promoter driving expression of firefly luciferase and the ␤-globin promoter driving expression of
Renilla luciferase. B, relative -fold activation of ␥-promoter/firefly as a
function of the amount of transfected plasmid encoding the transcription factors. C, the relative -fold activation of ␤-promoter/Renilla luciferase as functions of the amount of transfected plasmid encoding the
transcription factors. Each data point in B and C is the average of three
separate experiments, and the error bars (⫾1 S.D.) are indicated in the
graphs to show the variation between the experiments.
identify any sequences within the LCR or promoters with less
than a 7-base mismatch. Fuzznuc identified several sequences
with 7-base or more mismatches dispersed throughout the LCR
and promoters.
Regulation of Endogenous ␥-Globin Expression in Transfected Cells—Encouraged by the high level of activation of the
␥-globin promoter in the gg1-VP64-HA reporter assay, we analyzed transcription of the endogenous ␥-globin gene in the
presence of this TF. The erythroleukemia cell line K562 was
co-transfected with pcDNA-gg1-VP64-HA and pLEGFP, or
pcDNA3.1 (empty vector) and pLEGFP, followed by cultivation
for 96 h. The cells were subsequently sorted, and the 10% with
the highest enhanced green fluorescent protein values were
collected. Intracellular lysates from the collected cells were
analyzed for HbF content by Western blotting (Fig. 4A). Densitometric quantification of the bands in Fig. 4A showed 8-fold
higher content of HbF in cells transfected with the transcriptional activator than those transfected with the empty vector.
Chromosomal Integration of the Designed TF—To characterize the K562 cells expressing gg1-VP64-HA in more detail,
several clones with genomic integration of the pcDNAgg1-VP64-HA construct were generated. Genomic clones were
prepared by transfection of pcDNA-gg1-VP64-HA into K562
cells followed by selection with Geneticin G418. Individual
surviving clones were expanded. In order to quantify the -fold
of up-regulation of HbF in these clones, intracellular lysates
were prepared and analyzed for HbF levels by Western blotting
(Fig. 4b). The level of HbF in these clones varied from 7- to
16-fold higher than that of untransfected K562 cells as determined by densitometric quantification. One of these clones,
labeled D8, was used for further studies.
Analysis of the ␥-Globin and gg1-VP64-HA mRNA Levels in
Stably Transfected K562 Cells—To analyze the changes in
3712
Targeted Up-regulation of Human ␥-Globin
FIG. 4. Western blotting using an HbF-specific antibody. A, a
Western blotting of intracellular lysates of enhanced green fluorescent
protein-positive cells that were co-transfected with pLEGFP and either
pcDNA3.1 (lane 1) or pcDNA-gg1-VP64-HA (lane 2). Densitometric
quantification of the difference between the bands in lane 1 and 2 shows
an 8-fold increase. B, Western blotting of individual clones with an
integrated copy of pcDNA-gg1-VP64-HA. Lane 1, normal K562 cells;
lane 2, D8 clone (10-fold up-regulation); lane 3, E15 clone (7-fold upregulation); lane 4, C22 (16-fold up-regulation).
mRNA transcription, total RNA from normal K562 cells and
from the stably transfectant clone D8 was collected and converted to cDNA. Analysis for the presence of gg1-VP64-HA
mRNA was performed using these cDNA samples as template
in a nested RT-PCR with gg1-VP64-HA-specific primers. The
gg1-VP64-HA mRNA was indeed present in the D8 cells (Fig.
5A, lane 2). No gg1-VP64-HA mRNA was detected in the sample from the normal K562 cells (Fig. 5A, lane 1). An mRNA
sample from the D8 clone where no reverse transcriptase was
added during cDNA synthesis was also negative (Fig. 5A, lane
3). Hence, the band in Fig. 5A, lane 2, does not originate from
a genomic DNA template and is indeed the amplification of a
reverse transcribed mRNA. The relative amounts of ␥-globin
mRNA were determined by real time PCR using the cDNA
samples. This showed that the ␥-globin mRNA levels were 2.5
times higher in the D8 clone compared with native K562 cells.
Visual and Flow Cytometric Analysis of Retrovirally Transduced Cells and the D8 Cell Line—In order to further study
targeting and the ability of gg1-based transcription factors to
directly regulate the endogenous ␥-globin gene, gg1 was subcloned into pMX retroviral vectors and expressed as a fusion to
either the activation domain VP64 or the transcriptional repression domain Krüppel-associated box (KRAB) (29, 35). The
effect of gg1 transcription factors was considerable in retrovirally transduced K562 cells in both activation and repression of
␥-globin expression. Visual inspection of the cell pellet of retrovirally transduced cells and the cell pellet of native K562
cells showed that the cells transduced with the VP64 activator
were significantly more red-colored, whereas the cells transduced with the KRAB repressor were more opaque than the
native cells (Fig. 6A). The variation in red coloration of the cells
is probably due to differing levels of HbF. The cells were also
stained with an antibody specific for HbF and analyzed in a
flow cytometer (Fig. 6B), where the HbF level was reduced by
90% in the repressor-transduced cells and increased 14-fold
with VP64 activator-transduced cells (Fig. 6B).
In comparison, the D8 clonal line stably expressing the activator had a 8-fold higher amount of HbF than native K562 cells
as determined by fluorescence-activated cell sorting analysis
and also demonstrated a more intense red color in the cell
pellet. Fluorescence-activated cell sorting determination of
HbF content was comparable with the level of up-regulation
determined by Western blotting (10-fold). The D8 clone could
be cultivated for over 3 months without any apparent loss of
HbF expression, as judged by periodic visual inspection of
FIG. 5. A, analysis of RT-PCR performed with primers specific for the
gene encoding gg1-VP64-HA. Lane 1, a negative control, corresponding
to K562 cells. Lane 2, corresponds to the D8 clone with stable expression
of gg1-VP64-HA. Lane 3, a negative control of the D8 clone, where no RT
was added in the cDNA synthesis. B, RT-PCR performed with GADPHspecific primers on the same samples as a control.
pelleted cells and flow cytometric analysis.
In Vivo Binding of Transcription Factors as Determined by
Chromatin Immunoprecipitation—A chromatin immunoprecipitation (ChIP) experiment was performed to further establish the direct interaction of gg1 transcription factors with the
␥-globin promoter site in vivo (Fig. 6C). A chromatin fragment
comprising the ␥-globin promoter site was efficiently crosslinked by the gg1-VP64-HA and gg1-KRAB-HA proteins expressed in K562 cells and co-immunoprecipitated with a polyclonal antibody raised specifically against the framework of our
designer zinc finger domains (25). No immunoprecipitated
DNA fragment was detected by PCR using native K562 cells or
without the addition of the antibody. This demonstrates that
the immunoprecipitated complex was only formed with our
artificial transcription factors and not with endogenous zinc
finger-containing factors and that our designed transcription
factors interact with the ␥-globin promoter site directly in vivo.
Analysis of Expression of gg1-VP64-HA and HbF Using Deconvolution Fluorescence Microscopy—Individual cells from the
stably transfected clone D8, normal K562 cells, and K562 cells
treated with the HbF inducer hemin were studied in a deconvolution microscope. Cell samples were collected and permeabilized with saponin. This was followed by staining with
DAPI, an anti-HA monoclonal antibody, and an anti-HbF
monoclonal antibody, which interact with the cell nucleus, gg1VP64-HA, and HbF, respectively. The cells were subsequently
visualized in a deconvolution microscope, and the maximal
projections of three-dimensional datasets are shown in Fig. 7.
Cells of the D8 clone and of hemin-stimulated cells contained
an increased level of HbF compared with the K562 cells (Fig. 7,
compare g and h with i). Fig. 7 also shows that the zinc finger
was present only in the D8 clone (Fig. 7, compare e with d and
f). In addition, HbF was located in the cytoplasm of the cells,
and gg1-VP64-HA was located in the nucleus of the cells (Fig.
7, j, k, and l).
Targeted Up-regulation of Human ␥-Globin
3713
FIG. 7. A co-localization experiment was performed, where
cells were stained for the TF and also for HbF and analyzed in
a deconvolution microscope. K562 cells were stimulated with hemin
(a, d, g, and j) or stably expressed gg1-VP64-HA (b, e, h, and k).
Nontreated K562 cells were included as a negative control (c, f, i, and l).
The top row shows the nucleus in blue; the second row shows gg1VP64-HA in green; the third row shows HbF in red, and the bottom row
shows an overlay picture of the rows above.
FIG. 6. A, a pellet of normal K562 cells and K562 cells retrovirally
transduced with gg1-KRAB repressor or with gg1-VP64 activator. B, an
overlay of fluorescence-activated cell sorting profiles of K562 cells
stained with an anti-HbF-specific antibody. The blue peak corresponds
to cells that were stained with a nonspecific phycoerythrin-labeled IgG1
antibody (negative control). The black peak corresponds to stained K562
cells, which show the basal level of HbF. The green peak corresponds to
cells where HbF was repressed by gg1-KRAB. The red peak corresponds
to cells expressing gg1-VP64 with an up-regulated expression of HbF. C,
in vivo binding of retrovirally expressed gg1-VP64 and gg1-KRAB transcription factors to the ␥-globin promoter by ChIP in K562 cells. Chromatin immunoprecipitation performed with antibody (ZF) or without
antibody (No) was compared with input chromatin (In), shown as positive control. ChIP assays in native K562 cells or without DNA control
(⫺) revealed no immunoprecipitated DNA. No enrichment of immunoprecipitated chromatin was detected with PCR primers amplifying an
unspecific (ICAM-1) promoter used as negative control (25).
DISCUSSION
We have designed and analyzed three zinc finger-based transcription factors for transactivation of the ␥-globin gene. When
identifying possible target sites within the promoter for TF interaction, several issues were taken into consideration. One important issue was the distance of the TFs from the start of
transcription. We were aided in this case by considerable data on
sequences within the ␥-globin promoter where endogenous enhancers and silencers interact. For example, the two cis-acting
␥-globin activators FKLF and FKLF-2 interact with the ⫺145
CACCC-box (36, 37). This sequence has been shown to be important for high level activation of the ␥-globin promoter (38). The
⫺112 and ⫺85 CCAAT-boxes constitute a minimal promoter and
can direct transcription of a low level of the ␥-globin gene (39).
Mutations in either of the CCAAT-boxes decrease the activity of
the promoter (40). Also, a G to A mutation at position ⫺117 has
been shown to increase expression of ␥-globin in cases of hereditary persistence of fetal hemoglobin (41– 43). The molecular
mechanism responsible for HPFH (⫺117 G 3 A) remains unclear. However, a study analyzing DNA-protein interactions in
the ␥-globin promoter in vivo revealed that an unknown protein
interacts with the promoter in the CCAAT-box situated next to
position ⫺117 (44). It is possible that this DNA-protein interaction is affected by the mutation. We therefore reasoned that this
portion of the promoter would be accessible to artificial transcription factors and would be at a suitable distance from the start of
transcription. Another issue was target site accessibility. We
reasoned that the appearance of the ⫺145 CACCC-box and the
⫺112 and ⫺85 CCAAT-boxes in this region implied that the
sequences would be accessible to designed TFs.
We were able to show that the DNA-binding domains of each of
the three TFs constructed were able to bind to double-stranded
hairpin DNA corresponding to sequences in the ␥-globin promoter with low to subnanomolar affinity. When analyzed for the
ability to interact with the full ␥-globin promoter and induce
expression of luciferase in a reporter construct, however, only one
of them (gg1-VP64-HA) could induce expression above the level of
the controls. Furthermore, the activity of this transcription factor
was sensitive to mutations introduced into its targeted binding
site. It may be that gg2-VP64-HA and gg3-VP64-HA failed to
activate transcription, since their binding sites are not at the
correct distance from the start of transcription. We note that the
binding site of gg1-VP64-HA is closer to the CACCC-boxes where
FKLF and FKLF-2 bind than the sites of the inactive designed
TFs. The FKLF and FKLF-2 sites may have evolved to be at an
optimal distance from the start of transcription. Another reason
why gg1-VP64-HA activated transcription whereas the other TFs
Targeted Up-regulation of Human ␥-Globin
3714
did not may be that the affinity of gg2-VP64-HA (5.3 nM) and
gg3-VP64-HA (5.4 nM) for their respective target DNA sequences
was almost 10 times weaker that the affinity of gg1-VP64-HA for
its target sequence (0.7 nM). A correlation between transcription
factor affinity and activation or repression activity has been
noted in studies targeting the erbB-2 gene and the HIV-1 long
terminal repeat (20, 45).
We demonstrated significant induction of endogenous ␥-globin expression in the erythroleukemia cell line K562 upon
transfection or retroviral transduction of cells with gg1-VP64HA. To demonstrate that the regulation of ␥-globin expression
by gg1-VP64-HA was direct, we examined the ability of gg1KRAB, a designed repressor bearing the same DNA recognition
domain, to repress transcription of the same gene. We demonstrated not only activation of ␥-globin with the retrovirally
delivered activator but also near complete repression of ␥-globin expression upon delivery of the repressor to K562 cells,
providing strong evidence for direct target ␥-globin gene regulation as opposed to nonspecific binding of the transcription
factor that might lead to erythroid differentiation of K562 cells
and subsequent indirect up-regulation of ␥-globin gene expression. Furthermore, the direct in vivo interaction of our designed
transcription factors with the ␥-globin promoter was demonstrated in chromatin immunoprecipitation experiments.
The level of HbF increased 7–16-fold in K562 cells stably
transfected or retrovirally transduced with gg1-VP64-HA. This
level of induction is higher than values reported by others in
studies of hydroxyurea-treated K562 cells (46). It is not presently known whether the level of HbF increase in K562 cells
induced by the transcription factor gg1-VP64-HA will be sufficient to ameliorate symptoms of SCD or thalassemic diseases
in human patients. Studies performed on patients suffering
from SCD have shown that normal levels of HbF vary greatly
within the population (from 2–32%) (47). However, major organ
failure, stroke, and aseptic necrosis, all leading to premature
death, seem to be ameliorated at levels of HbF higher than 10%
(47), and an even higher corrective effect is seen at HbF levels
higher than 20%, which significantly reduces SCD caused hospitalization (48). Thus, we are hopeful that our designed transcription factor could have a favorable effect on SCD patients.
There are several potential ways of delivering this TF to
affected patients. One strategy could include collection of hematopoietic stem cells from the patient affected by SCD or a
thalassemic disease followed by retroviral infection with virions carrying the gene encoding gg1-VP64-HA. Transduced cells
could then be delivered back to the individual. This strategy
may overcome problems with graft versus host disease and
difficulties with finding a suitable donor, two major obstacles
associated with hematopoietic stem cell transplantation. The
potential use of zinc finger-based TF therapeutics agents has
been demonstrated in murine models (22, 49). We expect that
the TFs described in this paper will be nontoxic, since the cells
maintain the same doubling time when expressing the TF (data
not shown). By utilizing information regarding the binding of
proximal native transcription factors coupled with genetic information, we have been able to design a transcription factor
that efficiently up-regulates transcription of endogenous ␥-globin in the erythroleukemia cell line K562, leading to an increased content of HbF in the cells. This lays the foundation for
experiments in disease systems of SCD and thalassemia in
animal models as well as gene therapy experiments with patients suffering from these diseases.
REFERENCES
1. Kollias, G., Wrighton, N., Hurst, J., and Grosveld, F. (1986) Cell 46, 89 –94
2. Chada, K., Magram, J., Raphael, K., Radice, G., Lacy, E., and Costantini, F.
(1985) Nature 314, 377–380
3. Chada, K., Magram, J., and Costantini, F. (1986) Nature 319, 685– 689
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
Rutherford, T., and Nienhuis, A. W. (1987) Mol. Cell. Biol. 7, 398 – 402
Peterson, K. R. (2003) Curr. Opin. Hematol. 10, 123–129
Wijgerde, M., Grosveld, F., and Fraser, P. (1995) Nature 377, 209 –213
Li, Q., Peterson, K. R., Fang, X., and Stamatoyannopoulos, G. (2002) Blood
100, 3077–3086
Stamatoyannopoulos, G., and Grosveld, F. (2001) in The Molecular Basis of
Blood Diseases (Stamatoyannopoulos, G., Majerus, P., Perlmutter, R., and
Varmus, H., eds) pp. 135–182, W. B. Saunders Co., Philadelphia, PA
Murray, N., Serjeant, B. E., and Serjeant, G. R. (1988) Br. J. Haematol. 69,
89 –92
Swank, R. A., and Stamatoyannopoulos, G. (1998) Curr. Opin. Genet. Dev. 8,
366 –370
Blouin, M.-J., Beauchemin, H., Wright, A., De Paepe, M., Sorette, M., Bleau,
A.-M., Nakamoto, B., Ou, C.-N., Stamatoyannopoulos, G., and Trudel, M.
(2000) Nat. Med. 6, 177–182
May, C., Rivella, S., Callegari, J., Heller, G., Gaensler, K. M. L., Luzzatto, L.,
and Sadelain, M. (2000) Nature 406, 82– 86
Desjarlais, J. R., and Berg, J. M. (1992) Proc. Natl. Acad. Sci. U. S. A. 89,
7345–7349
Wu, H., Yang, W. P., and Barbas, C. F., III (1995) Proc. Natl. Acad. Sci. U. S. A.
92, 344 –348
Greisman, H. A., and Pabo, C. O. (1997) Science 275, 657– 661
Segal, D. J., Dreier, B., Beerli, R. R., and Barbas, C. F., III (1999) Proc. Natl.
Acad. Sci. U. S. A. 96, 2758 –2763
Dreier, B., Beerli, R. R., Segal, D. J., Flippin, J. D., and Barbas, C. F., III (2001)
J. Biol. Chem. 276, 29466 –29478
Dreier, B., Segal, D. J., and Barbas, C. F., III (2000) J. Mol. Biol. 303, 489 –502
Seipel, K., Georgiev, O., and Schaffner, W. (1992) EMBO J. 11, 4961– 4968
Beerli, R. R., Dreier, B., and Barbas, C. F., III (2000) Proc. Natl. Acad. Sci.
U. S. A. 97, 1495–1500
Tan, S., Guschin, D., Davalos, A., Lee, Y. L., Snowden, A. W., Jouvenot, Y.,
Zhang, H. S., Howes, K., McNamara, A. R., Lai, A., Ullman, C., Reynolds,
L., Moore, M., Isalan, M., Berg, L. P., Campos, B., Qi, H., Spratt, S. K., Case,
C. C., Pabo, C. O., Campisi, J., and Gregory, P. D. (2003) Proc. Natl. Acad.
Sci. U. S. A. 100, 11997–12002
Rebar, E. J., Huang, Y., Hickey, R., Nath, A. K., Meoli, D., Nath, S., Chen, B.,
Xu, L., Liang, Y., Jamieson, A. C., Zhang, L., Spratt, S. K., Case, C. C.,
Wolffe, A., and Giordano, F. J. (2002) Nat. Med. 8, 1427–1432
Xu, D., Ye, D., Fisher, M., and Juliano, R. L. (2002) J. Pharmacol. Exp. Ther.
302, 963–971
Beerli, R. R., and Barbas III, C. F. (2002) Nat. Biotechnol. 20, 135–141
Magnenat, L., Blancafort, P., and Barbas, C. F., III (2004) J. Mol. Biol. 341,
635– 649
Zhang, L., Spratt, S. K., Liu, Q., Johnstone, B., Qi, H., Raschke, E. E.,
Jamieson, A. C., Rebar, E. J., Wolffe, A. P., and Case, C. C. (2000) J. Biol.
Chem. 275, 33850 –33860
Liu, P. Q., Rebar, E. J., Zhang, L., Liu, Q., Jamieson, A. C., Liang, Y., Qi, H.,
Li, P. X., Chen, B., Mendel, M. C., Zhong, X., Lee, Y. L., Eisenberg, S. P.,
Spratt, S. K., Case, C. C., and Wolffe, A. P. (2001) J. Biol. Chem. 276,
11323–11334
Mannervik, M., Nibu, Y., Zhang, H., and Levine, M. (1999) Science 284,
606 – 609
Beerli, R. R., Segal, D. J., Dreier, B., and Barbas, C. F., III (1998) Proc. Natl.
Acad. Sci. U. S. A. 95, 14628 –14633
Segal, D. J., Beerli, R. R., Blancafort, P., Dreier, B., Effertz, K., Huber, A.,
Koksch, B., Lund, C. V., Magnenat, L., Valente, D., and Barbas, C. F., 3rd.
(2003) Biochemistry 42, 2137–2148
Laniel, M. A., Beliveau, A., and Guerin, S. L. (2001) Methods Mol. Biol. 148,
13–30
Skarpidi, E., Vassilopoulos, G., Li, Q., and Stamatoyannopoulos, G. (2000)
Blood 96, 321–326
Boyd, K. E., Wells, J., Gutman, J., Bartley, S. M., and Farnham, P. J. (1998)
Proc. Natl. Acad. Sci. U. S. A. 95, 13887–13892
Collis, P., Antoniou, M., and Grosveld, F. (1990) EMBO J. 9, 233–240
Margolin, J. F., Friedman, J. R., Meyer, W. K., Vissing, H., Thiesen, H. J., and
Rauscher, F. J., III (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 4509 – 4513
Asano, H., Li, X. S., and Stamatoyannopoulos, G. (1999) Mol. Cell. Biol. 19,
3571–3579
Asano, H., Li, X. S., and Stamatoyannopoulos, G. (2000) Blood 95, 3578 –3584
Gumucio, D. L., Rood, K. L., Blanchard-McQuate, K. L., Gray, T. A., Saulino,
A., and Collins, F. S. (1991) Blood 78, 1853–1863
Anagnou, N. P., Karlsson, S., Moulton, A. D., Keller, G., and Nienhuis, A. W.
(1986) EMBO J. 5, 121–126
Langdon, S. D., and Kaufman, R. E. (1998) Blood 91, 309 –318
Collins, F. S., Metherall, J. E., Yamakawa, M., Pan, J., Weissman, S. M., and
Forget, B. G. (1985) Nature 313, 325–326
Gelinas, R., Endlich, B., Pfeiffer, C., Yagi, M., and Stamatoyannopoulos, G.
(1985) Nature 313, 323–325
Berry, M., Grosveld, F., and Dillon, N. (1992) Nature 358, 499 –502
Ikuta, T., and Kan, Y. W. (1991) Proc. Natl. Acad. Sci. U. S. A. 88,
10188 –10192
Segal, D. J., Goncalves, J., Eberhardy, S., Swan, C. H., Torbett, B. E., Li, X.,
and Barbas, C. F., III (2004) J. Biol. Chem. 279, 14509 –14519
Haley, J. D., Smith, D. E., Schwedes, J., Brennan, R., Pearce, C., Moore, C.,
Wang, F., Petti, F., Grosveld, F., Jane, S. M., Noguchi, C. T., and Schechter,
A. N. (2003) Biochem. Pharmacol. 66, 1755–1768
Powars, D. R., Weiss, J. N., Chan, L. S., and Schroeder, W. A. (1984) Blood 63,
921–926
Powars, D. R., Chan, L., and Schroeder, W. A. (1989) Ann. N. Y. Acad. Sci. 565,
262–278
Xu, L., Zerby, D., Huang, Y., Ji, H., Nyanguile, O. F., de los Angeles, J. E., and
Kadan, M. J. (2001) Mol. Ther. 3, 262–273