Moss Tom. DNA-protein interactions: principles and protocols

Подождите немного. Документ загружается.

4Stockley

4. Filtration manifold and vacuum pump: We use a Millipore 1225 Sampling Mani-

fold (cat. no. XX27 025 50), which has 12 sample ports.

5. Liquid scintillation counter, vials, and scintillation fluid.

6. Siliconized glass test tubes.

7. TBE buffer: 89 mM Tris, 89 mM boric acid, 10 mM EDTA, pH 8.3.

8. Formamide/dyes loading buffer: 80% v/v formamide, 0.5X TBE, 0.1% w/v

xylene cyanol, 0.1% w/v bromophenol blue.

9. Sequencing gel electrophoresis solutions and materials: 19% w/v acrylamide,

1% w/v bis-acrylamide, 50% w/v urea in TBE.

10. Acetic acid (10% v/v).

3. Methods

3.1. Preparation of End-Labeled DNA

1. Digest the plasmid ( 20 µg in 200 µL) with the restriction enzymes used to release

a suitably sized DNA fragment (usually <200 bp). Extract the digest with an

equal volume of buffered phenol and add 2.5 volumes of ethanol to the aqueous

layer in order to precipitate the digested DNA. If preparing samples for inter-

ference assays) only one restriction digest should be carried out at this stage,

see Chapter 15.

2. Add 50 µL 1X CIAP reaction buffer to the ethanol-precipitated DNA pellet (<50 µg).

Add 1 U CIAP and incubate at 37°C for 30 min followed by the addition of a

further aliquot of enzyme and incubate for a further 30 min. Terminate the reac-

tion by adding SDS and EDTA to 0.1% (w/v) and 20 mM, respectively in a final

volume of 200 µL and incubate at 65°C for 15 min. Extract the digest with buff-

ered phenol, then with 1:1 phenol:chloroform, and, finally, ethanol precipitate

the DNA from the aqueous phase as above.

3. Redissolve the DNA pellet in 18 µL 1X T

PNK buffer. Add 20 µCi γ-[

P]-ATP

and 10 U T4 PNK and incubate at 37°C for 30 min. Terminate the reaction by

phenol extraction and ethanol precipitation (samples for interference assays

should be digested with the second restriction enzyme at this poin)t. Redissolve

the pellet in nondenaturing gel loading buffer and electrophorese on a non-

denaturing polyacrylamide gel.

4. After electrophoresis, separate the gel plates, taking care to keep the gel on the

larger plate. Cover the gel with plastic wrap and in the darkroom, under the safe-

light, tape a piece of X-ray film to the gel covering the sample lanes. With a syringe

needle, puncture both the film and the gel with a series of registration holes. Alter-

natively, register the film and the gel using fluorescent marker strips. Locate the

required DNA fragments by autoradiography of the wet gel at room temperature

for several min (approx 10 min). Excise slices of the gel containing the bands of

interest using the autoradiograph as a guide. Elute the DNA into elution buffer

overnight (at least) at 37°C. Ethanol precipitate the eluted DNA by adding 2.5 vol

of ethanol, wash the pellet thoroughly with 70% v/v ethanol, dry briefly under

vacuum, and rehydrate in a small volume (approx 50 µL) of TE. Determine the

radioactivity of the sample by liquid scintillation counting of a 1-µL aliquot.

Filter-Binding Assays 5

3.2. Filter-Binding Assays

3.2.1. Determination of the Equilibrium Constant

1. Presoak the filters in FB at 4°C for several hours before use. Care must be taken

to ensure that the filters are completely “wetted.” This is best observed by laying

the dry filters carefully onto the surface of the FB using blunt-ended tweezers

and observing buffer uptake.

2. Prepare a stock solution of radioactively labeled DNA fragment in an appropri-

ate buffer, such as FB. We adjust conditions so that each sample to be filtered

contains roughly 20 kcpm. Under these conditions, the DNA concentration is

<1 pM. Aliquot the stock DNA solution into plastic Eppendorf tubes. It is best at

this stage if relatively large volumes are transferred in order to minimize errors

caused by pipeting. We use 180 µL/sample. If the DNA-binding protein being

studied requires a cofactor, it is best to add it to the stock solution at saturating

levels so that its concentration is identical for every sample.

3. Prepare a serially diluted range of protein concentrations diluting into BB. A

convenient range of concentrations for the initial assay is between 10

–11

and

–5

M protein.

4. Immediately add 20 µL of each protein concentration carefully to the sides of the

appropriately labeled tubes of stock DNA solution. When the additions are com-

plete centrifuge briefly (5 s) to mix the samples and then incubate at a tempera-

ture at which complex formation can be observed (37°C for MetJ). For each

binding curve it is important to prepare two control samples. The first contains

no protein in the 20 µL of BB and is filtered to determine the level of background

retention. The second is identical to the first but is added to a presoaked filter in

a scintillation vial (see step 6) and is dried directly without filtering. This gives a

value for 100% input DNA.

5. After an appropriate time interval to allow equilibrium to be established, recen-

trifuge the tubes to return the liquid to the bottom of the tube and begin filtering.

6. The presoaked filters are placed carefully on the filtration manifold ensuring that

excess FB is removed and that the filter is not damaged. Cracks and holes are

easily produced by rough handling. The sample aliquot (200 µL) is then immedi-

ately applied to the filter, where it should be held stably by surface tension. Apply

the vacuum. If further washes are used they should be applied as soon as the

sample volume has passed through the filter. Remove the filter to a scintillation

vial and continue until all the samples have been filtered.

7. The scintillation vials should be transferred to an oven at 60°C to dry the filters

thoroughly (approx 20 min) before being allowed to cool to room temperature

and 3–5 mL of scintillation fluid added. The radioactivity associated with each

filter can now be determined by counting on an open channel (see Note 2).

8. Correct the value for each sample by subtracting the counts in the background

sample (no protein). Calculate the percentage of input DNA retained at each pro-

tein concentration using the value for 100% input from the unaltered sample. Plot

a graph of percentage retained vs the logarithm of the protein concentration (e.g.,

6Stockley

Fig. 1). The binding curve should increase from left to right until a plateau is

reached. This is rarely at 100% of input DNA. The plateau value can be assumed

to represent the retention efficiency, and for quantitative measurements, the

data points can be adjusted accordingly. There is not enough space here to

describe in detail the form of the binding curve or how best to interpret the

data. (For an authoritative yet accessible account, see ref. 9). For our pur-

poses, the protein concentration at 50% saturation can be thought of as the equi-

librium dissociation constant.

9. Once an initial binding curve has been obtained, the experiment should be

repeated with sample points concentrated in the appropriate region (i.e., the

region where the percentage retained is changing most rapidly).

Control experiments with DNAs that do not contain specific binding sites

should also be carried out to prove that binding is sequence-specific. Highly

diluted protein solutions appear to lose activity in our hands, possibly because of

nonspecific absorption to the sides of tubes, among other things. We therefore

produce freshly diluted samples daily. BB can be stored at 4°C for several days

without deleterious effect. Ideally, binding curves should be reproducible. How-

ever, there is some variability between batches of filters and we therefore recom-

mend not switching lot numbers during the course of one set of experiments.

3.2.2. Kinetic Measurements

Kinetic analysis of the binding reaction depends on prior determination of the

equilibrium binding curve, especially the concentration of DNA-binding protein

required to saturate the input DNA. This information allows a reaction mixture

containing a limiting amount of protein to be set up (e.g., at a protein concentration

that produces 75% retention). Both association and dissociation kinetics can be

studied. The major technical problem arises because of the relatively rapid sam-

pling rates that are required. However, it is almost always possible to adjust solu-

tion conditions such that sampling at 10 s intervals is all that is needed. Dissociation

measurements often need to be made over periods of up to 1 h, whereas association

reactions are usually complete within several min.

3.2.2.1. DISSOCIATION

Repeat steps 1 and 2 of Subheading 3.2.1. but do not aliquot the stock DNA

solution. Add to this sample the appropriate concentration (i.e., which pro-

duces approx 75% retention) of stock protein and allow to equilibrate. Add a

20-fold excess of unlabeled DNA fragment containing the binding site and

begin sampling (approx 200 µL aliquots) by filtration. Plots of radioactivity

retained vs time can then be analyzed to derive kinetic constants. In the sim-

plest case of a bimolecular reaction, a plot of the natural logarithm of the radio-

activity retained at time t divided by the initial radioactivity vs time yields the

first-order dissociation constant from the slope. An important control experi-

Filter-Binding Assays 7

ment is to repeat the experiment with DNA that does not contain a specific

binding site to show that dissociation is sequence-specific.

A variation of this experiment can be used in which the concentration of

protein in the reaction mix is diluted across the range where most complex

formation occurs. In this case it is necessary to prepare the initial complex in a

small volume (approx 50 µL) and then dilute 100 times with BB, followed by

filtering 500 µL aliquots.

3.2.2.2. ASSOCIATION

Set up a stock DNA concentration in a single test tube in (Subheading 3.2.1.

(steps 1 and 2). Incubate both this DNA and the appropriate solution of protein

at the temperature at which complexes form. Add the appropriate volume of

protein (e.g., 200 µL) to the DNA stock solution (1800 µL) and immediately

begin sampling (10 × 200 µL aliquots).

3.2.3. Interference Measurements

Experiments of this type can be used to gain information about the site on

the DNA fragment being recognized by the protein. The principle is identical

to that used in gel retardation interference assays but has the advantage that the

DNA does not have to be eluted from gels after fractionation.

1. Modify the purified DNA fragment radiolabeled (approx 100 kcpm) at a single

site with the desired reagent; for example, hydroxyl radicals, which result in the

elimination of individual nucleotide groups (10) (see Chapter 16), dimethyl sul-

fate (DMS) (11) (see Chapter 14), which modifies principally guanines, or ethyl

nitrosourea, ENU (see Chapter 15), which ethylates the nonesterified phosphate

oxygens. The extent of modification should be adjusted so that any one fragment

has no more than one such modification. This can be assessed separately in test

reactions and monitored on DNA sequencing gels.

2. Ethanol precipitate the modified DNA, wash twice with 70% (v/v) ethanol and

then dry briefly under vacuum. Resuspend in 200 µL FB. Remove 20 µL as a

control sample. Add 20 µL of the appropriate protein concentration to form a

complex and allow equilibrium to be reached. Filter as usual but with a siliconized

glass test tube positioned to collect the filtrate. (The Millipore manifold has an

insert for just this purpose.) Do not over dry the filter.

3. Place the filter in an Eppendorf tube containing 250 µL FB, 250 µL H

O, and

0.5% (w/v) SDS. Transfer the filtrate into a similar tube and then add SDS

and H

O to make the final volume and concentration the same as the filter-

retained sample. Add an equal volume of buffer-saturated phenol to each tube,

vortex, and centrifuge to separate the phases. Remove the aqueous top layers,

re-extract with chloroform:phenol (1:1), and then ethanol precipitate. A Geiger

counter can be used to monitor efficient elution of radioactivity from the filter,

which can be re-extracted if necessary.

8Stockley

4. Recover all three DNA samples (control, filter-retained, and filtrate) after etha-

nol precipitation and, if necessary, process the modification to completion (e.g.,

piperidine for DMS modification, NaOH for ENU, and so on). Ethanol precipi-

tate the DNA, dry briefly under vacuum, and then redissolve the pellets in 4 µL

formamide/dyes denaturing loading buffer. At this stage, it is often advisable to

quantitate the radioactivity in each sample by liquid scintillation counting of

1-µL aliquots. Samples for sequencing gels should be adjusted to contain roughly

equal numbers of counts in all three samples.

5. Heat the samples to 90°C for 2 min and load onto a 12% w/v polyacrylamide

sequencing gel alongside Maxam–Gilbert sequencing reaction markers (12).

Electrophorese at a voltage that will warm the plates to around 50°C. After elec-

trophoresis, fix the gel in 1 L 10% v/v acetic acid for 15 min. Transfer the gel to

3MM paper and dry under vacuum at 80°C for 60 min. Autoradiograph the gel at

–70°C with an intensifying screen.

6. Compare lanes corresponding to bound, free, and control DNAs for differences

in intensity of bands at each position (see Note 3). A dark band in the “free frac-

tion” (and a corresponding reduction in the intensity of the band in the “bound

fraction”) indicates a site where prior modification interferes with complex for-

mation. This is interpreted as meaning that this residue is contacted by the pro-

tein or a portion of the protein comes close to the DNA at this point. (See Chapters

14–16 for more extensive discussions of interference experiments.)

3.3. Results and Discussion

Figure 1 shows a typical filter-binding curve for the E. coli methionine

repressor binding to its idealized operator site of (dAGACGTCT)

cloned into

a pUC-polylinker. In the presence of saturating amounts of cofactor (SAM), a

sigmoidal binding curve is produced, whereas in the absence of SAM, the bind-

ing curve does not saturate in the protein concentration range tested. Similar

binding curves have been analyzed to produce Scatchard and Hill plots (9) in

order to examine the cooperativity with respect to protein concentration (6).

However, such multiple binding events should also be studied by gel retardation

assays which yield data about the individual complex species (see Chapter 2).

Table 1 shows the results obtained for binding to a series of variant operator

sites and illustrates the apparent sensitivity of the technique. However, in order

to make such comparisons, it is essential to determine the binding curves accu-

rately and with the same batches of protein and filters to minimize minor dif-

ferences between experiments. Table 1 lists the affinities of a number of variant

met operator sites cloned into pUC-polylinkers as determined by filter binding

in the presence of saturating levels of corepressor, SAM. The repressor binds

cooperatively to tandem arrays of an 8-bp met-box sequence (dAGACGTCT)

with a stoichiometry of one repressor dimer per met-box. The variant operators

were designed to examine both the tandem binding and the alignment of

repressor dimers with the two distinct dyads in tandem met-box sequences (6).

Filter-Binding Assays 9

Operator variants are as follows:

1. 00045-A single 8-bp met-box or half-site. The binding curve does not saturate

because singly bound repressor dimers dissociate very rapidly.

2. 00048-Two perfect met-boxes representing the idealized minimum operator

sequence. Repressors bind cooperatively with high affinity.

3. 00184-Two met-boxes with the central T–A step reversed. The crystal structure

of the repressor–operator complex shows that the central T–A step is not con-

tacted directly by the repressors, rather the pyrimidine–purine step promotes a

sequence-dependent DNA distortion that results in protein–DNA contacts else-

where in the operator fragment. The A–T step has less tendency to undergo this

conformational change and this is reflected in its lowered affinity.

4. 00299-A “shifted” two met-box operator used to define the alignment between

the repressor twofold axis and the operator dyads. The low affinity of this con-

struct compared to 00048 confirms that each repressor dimer is centered on the

middle of a met-box.

3.3.1. In Vitro Selection Experiments

In recent years, in vitro selection experiments have been used to identify the

range of preferred DNA target sequences by DNA-binding proteins (13,14),

(see also Chapter 42). The technique depends on the separation of protein-

bound DNA sequences from unbound, nonspecific, or low-affinity sites. Filter

binding is an attractive option for this selection step because of the speed with

which filtration and recovery of the bound fraction can be achieved. However,

it is important to be aware that some minor DNA variants can be retained spe-

cifically by the filters, thus biasing the selected sequences. One way to avoid

this and still retain the advantages of filter binding is to alternate rounds of

filter binding with separation by gel retardation (see Chapter 2). A detailed

discussion of the factors involved in such experiments is beyond the scope

Table 1

The Relative

s of a Number of Variant Met Operator Sites

Variant Operator Sequence Relative K

00045 AGACGTCT >12.2

00048 AGACGTCTAGACGTCT 1.0

00184 AGACGTCatGACGTCT 2.8

00299 gtctAGACGTCTagac 4.9

Note: The K

is the concentration of protein that produces 50% binding of

input DNA. Values are averages of several experiments and are quoted relative

to the two met-box perfect consensus sequences (00048) which, under the con-

ditions used, had an apparent K

of 82 ± 5 nM MetJ monomer. Sequences in

capitals represent matches to the consensus met box.

10 Stockley

of this chapter and the reader is referred to more detailed descriptions (e.g.,

ref. 15).

4. Notes

1. None of the radioactivity is retained by the filter. This again can be caused by a

variety of factors. Check that the preparation of DNA-binding protein is still func-

tional (if other assays are available) or that the protein is still intact by SDS-

polyacrylamide gel electrophoresis. Check the activity/concentration of the

cofactor if required. A common problem we have encountered arises because of

the different grades of commercially available BSA. It is always advisable to use

a preparation that explicitly claims to be nuclease and protease free.

2. All of the radioactivity is retained by the filter. This is a typical problem when

first characterizing a system by filter binding and can have many causes. Check

that the filters being used “wet” completely in FB and do not dry significantly

before filtration. Make sure that the DNA remains soluble in the buffer being

used by simple centrifugation in a bench-top centrifuge. If the background

remains high, add dimethyl sulfoxide to the filtering solutions. Classically,

5% (v/v) is used but higher concentrations (approx 20% v/v) have been reported

with little, if any, effect on the binding reaction. We have experienced excessive

retention when attempting to analyze the effects of divalent metal ions on com-

plex formation, and, in general, it is best to avoid such buffer conditions.

3. Poor recoveries from the filter-retained samples in interference assays, or other

problems in processing such samples further, can often be alleviated by addition

of 20 µg of tRNA as a carrier during the SDS/phenol extraction step.

Acknowledgment

I am grateful to Yi-Yuan He for providing the data shown in Table 1 and

Fig. 1.

References

1. Nirenberg, M. and Leder, P. (1964) RNA codewords and protein synthesis. The

effect of trinucleotides upon the binding of sRNA to ribosomes. Science 145,

1399–1407.

2. Jones, O. W. and Berg, P. (1966) Studies on the binding of RNA polymerase to

polynucleotides. J. Mol. Biol. 22, 199–209.

3. Riggs, A. D., Bourgeois, S., Newby, R. F., and Cohn, M. (1968) DNA binding of

the lac repressor. J. Mol. Biol. 34, 365–368.

4. Riggs, A. D., Suzuki, H., and Bourgeois, S. (1970) lac repressor-operator interac-

tion. I. Equilibrium studies. J. Mol. Biol. 48, 67–83.

5. Riggs, A. D., Bourgeois, S., and Cohn, M. (1970) The lac repressor-operator

interaction. III. Kinetic studies. J. Mol. Biol. 53, 401–417.

6. Phillips, S. E. V., Manfield, I., Parsons, I., Davidson, B. E., Rafferty, J. B.,

Somers, W. S., et al. (1989) Cooperative tandem binding of Met repressor from

Escherichia coli. Nature 341, 711–715.

Filter-Binding Assays 11

7. Old, I. G., Phillips, S. E. V., Stockley, P. G., and Saint-Girons, I. (1991) Regula-

tion of methionine biosynthesis in the enterobacteriaceae. Prog. Biophys. Mol.

Biol. 56,145–185.

8. Ryan, P. C., Lu, M., and Draper, D. E. (1991) Recognition of the highly con-

served GTPase center of 23S ribosomal RNA by ribosomal protein L11 and the

antibiotic thiostrepton. J. Mol. Biol. 221, 1257–1268.

9. Wyman, J. and Gill, S. J. (1990) In Binding and Linkage: Functional Chemistry of

Biological Macromolecules, chap. 2, University Science Books, Mill Valley, CA.

10. Siebenlist, U. and Gilbert, W. (1980) Contacts between Escherichia coli RNA

polymerase and an early promoter of phage T7. Proc. Natl. Acad. Sci. USA 77,

122–126.

11. Hayes, J. J. and Tullius, T. D. (1989) The missing nucleoside experiment: a new

technique to study recognition of DNA by protein. Biochemistry 28, 9521–9527.

12. Maxam, A. M. and Gilbert, W. K. (1980) Sequencing end-labelled DNA with

base-specific chemical cleavages. Methods Enzymol. 65, 499–560.

13. Tuerk, C. and Gold, L. (1990) Systematic evolution of ligands by exponential

enrichment: RNA ligands to bacterophage T4 DNA polymerase. Science 249,

505–510.

14. Ellington, A. D. and Szostak, J. W. (1990) In vitro selection of RNA molecules

that bind specific ligands. Nature 346, 818–822.

15. Conrad, R. C., Giver, L., Tian, Y. and Ellington, A. D. (1996) In vitro selection of

nucleic acid aptamers that bind proteins. Methods Enzymol. 267, 336–367.

EMSAs for Analysis of DNA–Protein 13

From:

Methods in Molecular Biology, vol. 148: DNA–Protein Interactions: Principles and Protocols, 2nd ed.

Electrophoretic Mobility Shift Assays

for the Analysis of DNA-Protein Interactions

Marc-André Laniel, Alain Béliveau, and Sylvain L. Guérin

1. Introduction

Several nuclear mechanisms involve specific DNA–protein interactions. The

electrophoretic mobility shift assay (EMSA, also known as the gel mobility

shift or gel retardation assay), first described almost two decades ago (1,2),

provides a simple, efficient and widely used method to study such interactions.

Its ease of use, its versatility, and especially its high sensitivity (10

–18

mol of

DNA [2]) make it a powerful method that has been successfully used in a vari-

ety of situations not only in gene regulation analyzes but also in studies of

DNA replication, repair, and recombination. Although very useful for qualita-

tive purposes, EMSA has the added advantage of being suitable for quantita-

tive and kinetic analyzes (3). Furthermore, because of its very high sensitivity,

EMSA makes it possible to resolve complexes of different protein or DNA

stoichiometry (4) and even to detect conformational changes.

1.1. Principle of the Method

Electrophoretic mobility shift assay (EMSA) is based on the simple ratio-

nale that proteins of differing size, molecular weight, and charge will have

different electrophoretic mobilities in a nondenaturing gel matrix. In the case

of a DNA–protein complex, the presence of a given DNA-binding protein will

cause the DNA to migrate in a characteristic manner, usually more slowly than

the free DNA, and will thus cause a change or shift in the DNA mobility visible

upon detection.

While the kinetic analysis of EMSA, which has been extensively covered

elsewhere (ref. 5 and references therein), is not the prime focus of this chapter,

it will be useful to understand the basic theory underlying such analyzes. A

14 Laniel, Béliveau, and Guérin

univalent protein, P, binding to a unique site on a DNA molecule, D, will yield

a complex, PD, in equilibrium with the free components:

where k

is the rate of association and k

is the rate of dissociation. In the case

of a strong interaction between protein and DNA, with k

> k

, two distinct

bands are observed, corresponding to the complex PD and to the free DNA.

However, because of the dissociation that inevitably occurs during electro-

phoresis and because the DNA released from a complex during electrophoresis

can never catch up with the free DNA, a faint smear may be seen between the

two major bands. In contrast, a weak DNA–protein interaction, with k

< k

should produce a fainter band corresponding to the complex PD and a more

intense smear. However, even weak DNA–protein interactions may lead to

distinct bands in EMSA because of their stabilization in the gel matrix as a

result of the cage effect (6) and/or of molecular sequestration (7). In both cases,

the dissociation of the complex is slower within the gel than it is in free solu-

tion, but in the cage effect, the gel matrix prevents dissociated components P

and D from freely diffusing and thus favors a reformation of the complex PD,

whereas in molecular sequestration, the gel matrix isolates complex PD from

competing molecules that could promote its dissociation.

As for a single DNA molecule bearing multiple binding sites for a given

protein, there will generally be as many mobility shifts formed as there are

binding sites. For example, in the case of two independent binding sites on the

DNA fragment (D):

this would result in three DNA containing bands: the free DNA (D), the com-

plex with both sites occupied by protein (P2D), and the complexes with only

one occupied site (PD1 and PD2, which will generally migrate together).

The kinetics of more complex situations, such as dimerizing protein com-

plexes and multiple DNA–protein interactions, are beyond the scope of this

chapter, but some interesting and insightful articles have been recently pub-

lished (4,8) in which these questions are expressly addressed.

DPD

2P + DPD1 + P

PD 2 + P P2D