Moss Tom. DNA-protein interactions: principles and protocols
Подождите немного. Документ загружается.
420 Isalan and Choo
3.2. Construction of a Gene Cassette Coding
for a Protein Library
1. Design and synthesize the pool of related genes using end-to-end ligation of syn-
thetic “template” oligonucleotides, directed by annealing “guide” oligonucle-
otides, according to the schematic in Fig. 1. The DNA cassette must code for the
genes in frame with the leader sequence and geneIII. Note 3 provides guidelines
regarding the design of randomizations in the protein library.
2. Use polyacrylamide gel electrophoresis to purify about 100 mg of each oligonucle-
otide. Adjust purified oligonucleotide concentration to 10 pmol/µL in distilled water.
3. 5'-Phosphorylate the “template” oligonucleotides prior to ligation (Fig. 1). Take
10 µL of each oligonucleotide at 10 pmol/µL and add 1.2-µL of 10X T
4
ligase
buffer (this is essentially T
4
polynucleotide kinase buffer containing 1 mM ATP).
Add 1 µL (10 U) of T
4
polynucleotide kinase and incubate at 37°C for 1 h. Heat
inactivate the enzyme by incubating at 65°C for 10 min.
4. Take 100 pmol of each phosphorylated template oligonucleotide in (12-µL vol-
umes from kinase reactions) and mix with 100 pmol of each guide oligonucleotide
Fig 1. Construction of a gene cassette coding for the protein library to be displayed
on phage. The pool of genes is constructed by end-to-end ligation of synthetic “tem-
plate” oligonucleotides (T1–T3), containing nucleotide randomizations (marked X),
which is directed using “guide” oligonucleotides (G1, G2) that bridge the junctions by
sequence-specific annealing. After the ligation reaction, a full-length, double-stranded
gene cassette is amplified by the PCR using primers that contain restriction enzyme
sites for cloning.
Engineering Nucleic Acid-Binding Proteins 421
(10 µL of 10 pmol/µL in 1X T
4
DNA ligase buffer). Place in a boiling waterbath for
3 min, then turn off the heat source and leave to cool to room temperature to allow
annealing of the template and guide oligonucleotides. Finally, place on ice.
5. Add an equal volume of 1X T
4
DNA ligase buffer to the oligonucleotide mixture.
Supplement with 8 µL (3200 U [DNA end ligation units]) of T
4
DNA ligase per
100 µL. Incubate at 16°C for 8–24 h.
6. Carry out a small-scale (25-µL) trial polymerase chain reaction (PCR) reaction,
with appropriate primers (as shown in Fig. 1), to amplify full-length double-
stranded DNA cassette (see Note 4). Typically, 1 µL of the ligation reaction
contains sufficient template. The correct size of the PCR product should be veri-
fied by agarose gel electrophoresis. When PCR conditions are optimized, scale
up the PCR reaction 10-fold. If the PCR product is not of the predicted size, then
gel purification of the ligated single-stranded template may be required. Purifica-
tion is facilitated by
32
P end labeling of the 3'-terminal oligonucleotide used in
constructing the gene cassette.
7. Prepare the PCR product for restriction digestion by extracting once with phenol
and once with chloroform, followed by ethanol precipitation.
8. Resuspend DNA in 460 µL 1X NEBuffer 2 containing 100 µg/mL BSA. Add 10 µL
(200 U) of SfiI, overlay with mineral oil, and incubate at 50°C. Supplement the
reaction with 10 µL (200 U) of SfiI every 2 h to a total incubation time of 8 h.
9. Prepare DNA for the second digestion by extracting once with phenol and once
with chloroform, followed by ethanol precipitation.
10. Resuspend DNA in 460 µL 1X NEBuffer 3 containing 100 µg/mL BSA. Add 10 µL
(100 U) of NotI and incubate at 37°C. Supplement the reaction with 10 µL (200 U)
of NotI every 2 h to a total incubation time of 8 h.
11. Remove small restriction fragments and prepare the cassette for cloning using
agarose gel purification: run the cut cassette on 2% low-melting-point agarose
gel in 1X TAE containing 0.5 µg/mL ethidium bromide. Excise the band of interest
from the gel and extract DNA by any suitable method such as AgaraseI digestion.
12. The DNA yield should be quantitated and the cassette stored in distilled water at
–20°C. This protocol gives 1–3 µg of library cassette, of which less than 0.5 µg is
typically required to clone a library of 5 million transformants.
3.3. Cloning of the Library DNA Cassette into Phage Vector
1. Ligate DNA insert into a 1 µg (approx 0.2 pmol) phage vector. This is carried out
in a total volume of 30 µL in 1X T
4
DNA ligase buffer supplemented with 3 µL
(1200 U) of T
4
DNA ligase. Incubate at 16°C for 16 h. A 5:1 molar ratio of insert
to the vector should give good yields, but results can be improved by optimizing
this ratio after each new vector and insert preparation.
2. Purify ligation products by phenol/chloroform extraction followed by ethanol
precipitation. Wash the precipitate thoroughly with 70% ethanol and dry for 5 min
in a lyophilizer. Resuspend pellet in 10 µL of distilled water. It is important to
minimize the amount of salt carried through because this can cause arcing during
electroporation.
422 Isalan and Choo
3. Electroporate 2-µL (approx 200 ng) samples of ligated vector into 50–70 µL of
electrocompetent E. coli (see Note 5) in a 2-mm-path electroporation cuvet (Equibio).
4. Pulse cells in electroporation apparatus set to 2.5 kV and 25 µF, with a pulse
controller set to 200 Ω.
5. Immediately resuspend cells in 1 mL room-temperature SOC. Incubate for 1 h at
37°C with shaking.
6. Plate cells out on TYE medium containing 15 µg/mL tetracycline and grow 16 h
at 30°C. Large 24-cm × 24-cm plates are convenient for plating out approx 10
6
transformant colonies. Plate out a dilution series of the transformation on smaller
plates to estimate the total number of colonies obtained (i.e., the library size).
7. The efficiency of the cloning can be verified by PCR screening 20 randomly
selected colonies using internal primers that amplify the cloned cassette (see Note
4). Individual colonies can be cultured in liquid medium to produce phage (see
Note 6) for DNA sequencing to confirm that the library contains unbiased
randomizations at the expected positions.
8. Harvest the library by gently scraping colonies into 1–3 mL of 2X TY. Bacteria
may be stored at –70°C after the addition of 50% (v/v) glycerol.
9. To obtain phage particles, grow phage-infected bacteria in 2X TY containing
15 µg/mL tetracycline. Use 50 mL of culture medium/mL of harvested bacteria
and incubate for 8–24 h at 30°C.
10. Centrifuge cultures at 500g for 20 min to obtain clear phage-containing superna-
tant. This may be filtered through a 2-µm filter to remove bacteria, stored at 4°C
or –20°C, or alternatively used for selection experiments.
3.4. Phage Selection Against Nucleic Acid Targets
1. Nucleic acid targets used for selection are prepared by annealing complementary
DNA oligonucleotides together, or refolding an RNA hairpin structure. At least
one of the oligonucleotides must be biotinylated. Mix 10 µL of (each) oligo-
nucleotide at 10 pmol/µL with 20 µL of 2X nucleic acid annealing buffer. Place
in a boiling waterbath for 3 min, then turn off the heat source and allow to cool to
room temperature. Finally, place on ice. Annealed binding sites may be diluted in
water to a 1 pmol/µL stock solution stored at –20°C.
2. Fresh phage are prepared for selection by growing phage-infected bacteria in 2X
TY containing 15 µg/mL tetracycline for 8–24 h at 30°C. Grow 1 mL of culture
for each selection experiment (see Note 7).
3. Take 50 µL of dynabeads solution and separate the beads from the preservative
buffer using a magnet. Then, add 1 pmol of nucleic acid target site to the blocked
beads in 50 µL PBS and allow to bind for 30 min (see Note 8).
4. Start growing 1 mL of fresh TG1 bacteria per selection experiment, in 2X TY at
37°C. The bacteria will be ready when an OD
600
of 0.6 is reached (this takes
about 3 h if 2–5 mL of 2X TY is inoculated with a single colony). In the mean-
time, steps 5–10 of the selection should be performed.
5. Block the nucleic-acid-coated beads for 1 h at 20°C by adding 1 mL of PBS
containing 4% (w/v) fat-free freeze-dried milk (Marvel).
Engineering Nucleic Acid-Binding Proteins 423
6. Centrifuge bacterial cultures from step 2 on a bench-top microfuge at top speed
for 10 min to obtain clear phage-containing supernatants.
7. Prepare 1 mL phage binding mixture for each selection (see Note 9):
897 µL of PBS containing 2% (w/v) fat-free dried milk (Marvel) and 1%
(v/v) Tween-20
100 µL phage-containing supernatant (see Note 10)
2 µL (10 mg/mL) of sonicated salmon sperm DNA (or tRNA) competitor
1 mL (10 pmol) unbiotinylated competitor nucleic acid
8. Separate the nucleic-acid -coated beads from the blocking mixture and add 1 mL
of phage binding mixture. Incubate for up to 1 h at 20°C. RNA selections may
require shorter incubations at lower temperatures. We have selected variants of
the U1A RNA-binding motif by incubating for 5 min on ice. In this case, the
addition of ribonuclease inhibitors (e.g., RNasin [Promega]), is optional.
9. Wash away unbound phage from the beads with 15–20 washes of 1 mL PBS
containing 2% (w/v) fat-free freeze-dried milk (Marvel) and 1% (v/v) Tween-20.
Vortex the beads thoroughly during washes. The number of washes may need to
be optimized for a particular selection. If too few washes are carried out, then
selection is poor as many nonspecific binding phage are carried through. Too
many washes can reduce phage yield to the point where specific binders are lost.
Carry out one final wash with 1 mL PBS.
10. Remove PBS and elute phage from beads by adding 100 µL of 0.1 M triethanola-
mine. Mix on vortex for 1 min, then collect beads and remove the supernatant
that should be immediately neutralized with 100 µL of 1 M Tris buffer, pH 7.4.
11. Infect 500 µL of fresh logarithmic phase TG1 (from step 4) with 50 µL of eluted,
neutralized phage solution. Incubate without shaking at 37°C for 1 h. Dilutions of the
infected E. coli culture may be plated out on TYE containing tetracycline to estimate
phage yields in order to follow the progress of the selection over several rounds.
12. Centrifuge the infected culture for 5 min at top speed in a bench-top microfuge to
harvest the bacteria. Resuspend bacterial pellet in 2 mL 2X TY supplemented
with 15 µg/mL tetracycline, and grow for 16 h at 30°C to obtain phage for further
rounds of selection (see Note 11).
13. After the final round of selection, infected bacteria are plated out and individual
colonies are cultured to produce phage for binding assays and DNA sequencing.
3.5. Assaying Binding Properties of Selected Clones
by Phage ELISA
1. Prepare biotinylated nucleic acid binding sites as in step 1 of Subheading 3.4.
2. Prepare a fresh phage culture for ELISA (see Note 12) by inoculating 2 mL of
2X TY containing 15 µg/mL tetracycline with a single bacterial colony and incu-
bating for 8–24 h at 30°C.
3. Add between 0 and 100 pmol of biotinylated nucleic acid target site (in 50 µL of
PBS) to each streptavidin-coated microtiter well (250 µL capacity, Boehringer-
Mannheim) and allow to bind for 30 min. Initially, it may be worthwhile to carry
out binding assays using a range of nucleic acid target concentrations. Bear in
424 Isalan and Choo
mind the biotin binding capacity of the streptavidin-coated microtiter plate.
4. Block microtiter-plate wells for 1 h at 20°C by adding 150 µL PBS containing
4% (w/v) fat-free freeze-dried milk (Marvel).
5. Centrifuge phage cultures from step 2 on a bench-top microfuge for 10 min at top
speed to obtain clear phage-containing culture supernatant.
6. Prepare 50 µL phage-binding mixture for each ELISA well by mixing 5 µL phage
supernatant (see Note 10) with 45 µL of PBS containing 2% (w/v) fat-free freeze-
dried milk (Marvel), 1% (v/v) Tween-20, and 1-µg competitor nucleic acid (e.g.,
calf liver tRNA or sonicated salmon sperm DNA, depending on the application).
7. Discard blocking mixture from microtiter plate wells and add 50 µL of phage-
binding mixture per well. Incubate for up to 1 h at 20°C. Assays using RNA
targets may require shorter incubations at lower temperatures. We have assayed
variants of the U1A RNA-binding motif by incubating for 5 min on ice. The
addition of ribonuclease inhibitors (e.g., RNasin [Promega]) is optional.
8. Remove unbound phage by washing microtiter-plate wells seven times with PBS
containing 1% (v/v) Tween-20, followed by three washes with PBS.
9. Remove all PBS and add 100 µL of PBS containing 2% (w/v) fat-free freeze-
dried milk (Marvel) and 0.02% (v/v) peroxidase-conjugated anti-M13 IgG
(Pharmacia Biotech). Incubate for 1 h at 20°C.
10. About 10 min before the incubation from step 9 is over, prepare the reagents for
the ELISA developing solution. Dissolve a 1-mg TMB tablet in 100 µL of DMSO.
Make up a 10-mL solution of 0.1 M sodium acetate (pH 5.5) and add 2 µL of 30%
(v/v) hydrogen peroxide.
11. Remove unbound antibody by washing microtiter-plate wells three times with
PBS containing 0.05% (v/v) Tween-20, followed by three washes with PBS. Dis-
card all traces of PBS from the final wash.
12. Mix the solutions from step 10 to make developing solution and immediately add
100 µL to each microtiter-plate well.
13. Allow the colorimetric reaction to develop between 1–10 min. The time required
for developing depends on factors such as the concentration of nucleic acid target
bound to the plate, the amount of phage added to the reaction, and the quality of
the antibody used. A blue color develops in microtiter-plate wells that contain
phage, whereas empty control wells and wells to which phage have not bound
remain colorless.
14. Stop the reaction with 100 µL of 1 M sulfuric acid per well, which converts any
blue signal to yellow. Absorbance may be quantitated in a plate-reader spectro-
photometer at 450 nm.
4. Notes
1. Nucleic-acid-binding motifs such as the zinc finger (17) and the ribonucleopro-
tein motif (18) have been displayed on the surface of filamentous bacteriophage
as fusions to the minor coat protein encoded by gene III. This type of display can
be achieved using either phage vectors (which encode gIII and all functions
required for phage replication, packaging and infection, as well as antibiotic
Engineering Nucleic Acid-Binding Proteins 425
resistance) or phagemid vectors (which comprise only plasmid-encoded gIII and
require “rescue” with “helper phage” in order to replicate and package the single-
stranded genome) (19). Although either vector type is suitable for the experi-
ments described, each system has its own particular advantage. Phagemid vectors
are advantageous because they can be used to transform E. coli with greater effi-
ciency (10
7
–10
8
clones/µg DNA) relative to phage vectors (10
5
–10
6
clones/µg
DNA), hence facilitating cloning of larger combinatorial libraries. On the other
hand, we have found that phage vectors are more suitable for phage ELISA, which
can be used to study protein–nucleic acid interactions without the need to
subclone the selected genes into protein expression vectors. This may be because
the phage vectors permit polyvalent display of proteins, allowing for stronger
binding to ligands attached onto a solid support. Because the choice between
phage and phagemid depends on the application, we have almost always used a
phage vector, as we consider this vector system easier to work with and phage
ELISA to be particularly useful in our experiments.
2. Cesium chloride gradient purification of the phage vector is carried out as fol-
lows. Dissolve 10–100 µg of DNA in 4 mL TE buffer containing 4 g of cesium
chloride and 0.1 mL of ethidium bromide solution (10 mg/mL). Mix well to dis-
solve and dispense into a 5-mL ultracentifuge tube. Centrifuge at 370,000g for
20 h at 20°C using a Beckman Vti 65.2 rotor. Two bands of DNA should be
visible under UV light. The lower (supercoiled plasmid) band should be col-
lected with a syringe and washed four times with water-saturated butanol to
remove all traces of ethidium bromide. The resulting solution is diluted threefold
in TE buffer and then ethanol precipitated to recover pure phage vector.
3. A key step in designing the randomized gene cassette that will code for proteins
to be displayed on phage is to clearly define the purpose of the phage-display
experiment in order to decide the nature and extent of combinatorial randomiza-
tions. These decisions concern mainly (1) the part of the protein structure that is
to be varied, (2) the number of positions that should be randomized, and (3) the
amino acids to be represented at each varied position.
The process can be illustrated by analyzing the strategy behind our initial
experiments using the phage display of zinc fingers, which were designed to study
DNA recognition by the three-finger DNA-binding domain from transcription
factor Zif268. In the earliest experiment (8), our goal was to determine which
residue positions of the zinc-finger structure were responsible for DNA base rec-
ognition, and the amino acids at those positions which would effect nucleotide
discrimination. We therefore chose to randomize residues in only one zinc finger
of the three-finger DNA-binding domain such that the register of the protein–
DNA interaction would be defined by the interaction of the other two wild-type
fingers with the DNA-binding site.
The X-ray crystal structure of the Zif268 DNA-binding domain in complex
with DNA (20) showed that only three or four amino acids on the α-helix of each
zinc finger were responsible for base contacts. To determine whether it was
indeed these α-helical positions that were responsible for DNA-binding specific-
426 Isalan and Choo
ity, we decided to randomize several additional residues (in fact a total of seven
residues) in the zinc-finger α-helix in order to compare biases in the identity of
selected amino acids throughout the helix.
As we did not want to bias the selection to particular residues, we decided that
each randomization would represent most of the 20 amino acids, but we reduced
the theoretical library size by designing the randomized codons such that the first
base in the triplet was not thymine, thus precluding all stop codons and the codons
for Phe, Tyr, Trp, and Cys, which occur very rarely in the α-helices of zinc fin-
gers. The theoretical library size for the combinatorial variation of seven posi-
tions, each with 16 possible amino acids, was approx 5 × 10
8
; however, because
of the practical limitations (essentially the poor efficiency of transformation with
phage vector) we were only able to clone approx 2 × 10
6
members of the library.
Thus, the randomizations to be designed into a phage display library depend
partly on the aims of the experiment and partly on the practical limitations of the
technique, particularly the library size that can be cloned in practice.
Another important consideration in the phage display of proteins is that the
assembly of phage particles requires the gIII fusion protein to be translocated to
the periplasmic space of E. coli to await packaging (see ref. 15). Nor-
mally, this can be achieved by the presence of a cleavable signal peptide
sequence that is coded by most phage-display vectors. However, in the
event that phage yield is poor or that phage particles are not assembled at all,
particular attention must be paid to the amino acid sequence directly C-terminal
of the signal peptide. In many cases, it is worthwhile randomizing two to three
amino acids of this sequence in a pilot experiment, to select phage that are capable
of displaying the relevant nucleic-acid-binding protein. It should be noted, how-
ever, that certain nucleic acid binding motifs will prove very difficult or indeed
impossible to display on phage. In this case, the failure will become apparent
when phage particles fail to be produced on growing a culture of E. coli that have
been transformed with the recombinant phage vector.
4. Polymerase chain reactions typically contain 10 µL 10X Taq reaction buffer,
4 µL of each primer (stock conc., 10 pmol/µL), 1 µL of 20 mM dNTPs, 1.4 µL
Taq polymerase, and 1 µL ligation mixture, in a final volume of 100 mL.
5. Assuming all prior steps have proceeded smoothly, the efficiency of transforma-
tion will dictate the practical limit of the library size: 10
5
–10
6
transformants per
microgram of well-prepared phage vector is a typical yield. For best results, trans-
formations are carried out by electroporation of fresh (unfrozen) electrocompetent
cells prepared to a competency of at least 10
9
transformants per microgram
supercoiled pUC19 plasmid. As this technique requires practice and is also
unpredictable, it may be prudent to purchase such cells from a commercial supplier.
6. At this stage, it is worthwhile growing phage from individual colonies that con-
tain the recombinant phage vector in order to check that the construct is compe-
tent for phage particle assembly. The phage titer can be estimated by reinfecting
fresh E. coli from a liquid culture grown to mid-log phase, and plating on TYE
containing tetracycline.
Engineering Nucleic Acid-Binding Proteins 427
7. Certain nucleic-acid-binding motifs may require supplements to the growing
medium to ensure biological activity. For example, during phage display of zinc
fingers, we recommend supplementing all solutions with 50 µM zinc chloride.
8. The 1 pmol of target site will be used to select phage from 1 mL of binding
mixture. The concentration of binding site is therefore roughly in the nanomolar
range; phage that can bind with a K
d
below this range will be retained. For nucleic-
acid-binding motifs that bind their targets more weakly, the amount of target
site may have to be increased. Streptavidin-coated tubes or plates can also be
used for selection.
9. Phage displaying nucleic-acid-binding motifs of interest are selected from a large
library by affinity purification using biotinylated nucleic acid targets that have
been immobilized on a streptavidin-coated matrix. Unbound phage are washed
away and the phage that remain are eluted and amplified by passaging through
bacteria. After several (three to six) rounds of selection, the phage library
becomes enriched in those members that display proteins with the highest affini-
ties for the target sites.
Conditions of binding, washing, and elution can influence the results of a
phage-display experiment. We have found that selections using streptavidin-
coated paramagnetic beads (e.g., Dynabeads, Dynal) are more reliable than
selections using streptavidin-coated microtiter plates or tubes. The amount of
biotinylated nucleic acid target used for selection will determine the stringency
of the selection (see Note 8).
The amount and type of nonspecific competitor nucleic acid can also have an
effect. In early rounds of selection, nonspecific competitor such as sonicated
salmon sperm DNA, or tRNA, may be used. In later rounds, the selection may be
refined by adding competitor sequences that are closely related to the desired
target site but containing systematic base variations from the target sequence.
The composition of the binding and washing buffers will also affect the strin-
gency of selection. High-ionic-strength buffers tend to increase the stringency,
whereas relatively high concentrations of detergents help to reduce background.
It is important to wash away nonspecific binding phage thoroughly, and for this
purpose, we recommend at least 15 changes of buffer.
Finally, the method of elution can affect the results of the selection experi-
ments and can be manipulated accordingly; for example to yield all retained phage
(elution with alkali or acid) or only those bound phage with the highest specific-
ity (elution with nucleic competitors).
10. This protocol suggests a 1:10 dilution of the phage-containing supernatant, which
we find typically contains about 10
9
–10
10
colony-forming units/mL, but this is
only intended as a guideline. Depending on the application, we have also previ-
ously used 1:1 dilutions of the phage-containing supernatant, or indeed 50-fold
concentrates of the phage-containing supernatant. Higher concentrations of phage
are particularly useful during the early rounds of selection, when specifically
binding phage are rare in the library. Precipitation of the phage (e.g., by centrifu-
gation at 60,000 rpm for 20 min) is also useful prior to binding reactions involv-
428 Isalan and Choo
ing RNA, in order to minimize exposure to ribonucleases present in the bacterial
culture supernatant.
11. It is convenient to do one round of selection per day so that transfected phage can
be grown overnight for another round the next day. In practice, three to six rounds
of selection are usually sufficient to obtain clones of interest. Depending on the
nature of the library, however, the number of rounds required may vary greatly.
An alternative strategy is to plate out the bacteria infected with selected phage
after each round and to start a culture using a pool of the bacteria the next day.
This increases the number of phage that are produced in the overnight culture,
and reduces the selection advantage of fast-growing clones.
12. Phage clones isolated by selection against a nucleic acid target site must be
assayed to characterize their range of binding affinities and specificities. This is
conveniently done by phage ELISA, in which phage-displaying nucleic-acid-
binding proteins bind to microtiter-plate wells coated with appropriate nucleic-
acid-binding sites and are detected using an antiphage antibody. We routinely
perform phage ELISA using phage derived from Fd-TET-SN and displaying
either zinc-finger domains or the RNA-binding domain from U1A protein. How-
ever, we do not have any data on phage ELISA carried out with phagemid-derived
phage particles.
References
1. Smith, G. P. (1985) Filamentous fusion phage: novel expression vectors that dis-
play cloned antigens on the virion surface. Science 228, 1315–1317.
2. McCafferty, J., Griffiths, A. D., Winter, G., and Chiswell, D. J. (1990) Phage
antibodies: filamentous phage displaying antibody variable domains. Nature 348,
552–554.
3. Bass, S., Greene, R., and Wells, J. A. (1990) Hormone phage: an enrichment
method for variant proteins with altered binding properties. Proteins 8, 309–314.
4. Pabo, C. O. and Sauer, R. T. (1992) Transcription factors: structural families and
principles of DNA recognition. Annu. Rev. Biochem. 61, 1053–1095.
5. Klug, A. (1993) Protein designs for the specific recognition of DNA. Gene 135,
83–92.
6. Rebar, E. J. and Pabo, C. O. (1994) Zinc finger phage: affinity selection of fingers
with new DNA-binding specificities. Science 263, 671–673.
7. Jamieson, A. C., Kim, S.-H., and Wells, J. A. (1994) In vitro selection of zinc
fingers with altered DNA-binding specificity. Biochemistry 33, 5689–5695.
8. Choo, Y. and Klug, A. (1994) Toward a code for the interactions of zinc fingers
with DNA: Selection of randomised zinc fingers displayed on phage. Proc. Natl.
Acad. Sci. USA 91, 11,163–11,167.
9. Wu, H., Yang, W.-P., and Barbas III, C. F. (1995) Building zinc fingers by selec-
tion: Toward a therapeutic application. Proc. Natl. Acad. Sci. USA 92, 344–348.
10. Rebar, E. J., Greisman, H. A., and Pabo, C. O. (1996) Phage display methods for
selecting zinc finger proteins with novel DNA-binding specificities. Methods
Enzymol. 267, 129–149.
Engineering Nucleic Acid-Binding Proteins 429
11. Choo, Y. and Klug, A. (1995) Designing DNA-binding proteins on the surface of
filamentous phage. Curr. Opin. Biotech. 6, 431–436.
12. Friesen, W. J. and Darby, M. K. (1997) Phage display of RNA binding zinc fin-
gers from transcription factor IIIA. J. Biol. Chem. 272, 10,994–10,997.
13. Friesen, W. J. and Darby, M. K. (1998) Specific RNA binding proteins constructed
from zinc fingers. Nat. Struct. Biol. 5, 543–546.
14. Hoffmann, S. and Willbold, D. (1997) A selection system to study protein–RNA
interactions: functional display of HIV-1 Tat protein on filamentous bacterioph-
age M13. Biochem. Biophys. Res. Com. 235, 806–811.
15. Laird-Offringa, I. A. and Belasco, J. G. (1996) In vitro genetic analysis of RNA-
binding proteins using phage display libraries. Methods Enzymol. 267, 149–168.
16. Choo, Y. and Klug, A. (1994) Selection of DNA binding sites for zinc fingers
using rationally randomized DNA reveals coded interactions. Proc. Natl. Acad.
Sci. USA 91, 11,168–11,172.
17. Klug, A. and Rhodes, D. (1987) “Zinc fingers”: a novel protein motif for nucleic
acid recognition. Trends Biochem. Sci. 12, 464–469.
18. Nagai, K., Oubridge, C., Ito N, Avis J., and Evans P. (1995) The RNP domain: a
sequence-specific RNA-binding domain involved in processing and transport of
RNA. Trends Biochem. Sci. 20, 235–240.
19. Harrison, J. L., Williams, S. C., Winter, G., and Nissim, A. (1996) Screening of
phage antibody libraries. Methods Enzymol. 267, 83–109.
20. Pavletich, N. P. and Pabo, C. O. (1991) Zinc finger-DNA recognition: crystal
structure of a Zif268-DNA complex at 2.1 Å. Science 252, 809–817.