Moss Tom. DNA-protein interactions: principles and protocols

Подождите немного. Документ загружается.

160 Buckle and Travers

10. Rougée, M. and Bensasson, R. V. (1986) Détermination des constantes de vitesse

de désactivation de l’oxygène singulet (

) en presénce de biomolécules. Comp.

Rend. Acad. Sci. Paris 302, 1223–1226.

11. Travers, A. A., Lamond, A. I., Mace, H. A. F., and Berman, M. L. (1983) RNA

polymerase interactions with the upstream region of the E. coli tyrT promoters.

Cell 35, 265–273.

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

specific chemical cleavages. Methods Enzymol. 155, 560–568.

UV-Laser Footprinting 161

161

From:

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

Edited by: T. Moss © Humana Press Inc., Totowa, NJ

Ultraviolet-Laser Footprinting

Johannes Geiselmann and Frederic Boccard

1. Introduction

1.1. Measurement of DNA–Protein Interactions In Vitro

A large number of processes within the cell, in particular the regulation of

gene expression, rely on the binding of proteins to specific sites on the DNA. A

primary ingredient to understanding these processes is the characterization of

the protein–DNA interaction (1). Such a characterization consists in determin-

ing the position of the binding site on the DNA and measuring the affinity of

the protein for this recognition site. A wide variety of footprinting techniques

can accomplish this task (2,3).

1.2. Footprinting Techniques

These techniques involve the reaction of a footprinting reagent (in the larg-

est sense) with DNA and the subsequent localization and quantification of the

resultant DNA modification. Commonly used footprinting reagents include

DNase I, KMnO

, or dimethylsulfate (DMS) (2,3); see Chapters 3–14. Most of

these methods require extended incubation times of the reagent with the DNA–

protein complex, which may lead to artifacts if the footprinting reagent modi-

fies the complex. For example, DMS may react with the protein; DNase I

relaxes a supercoiled plasmid, which may destabilize the DNA–protein com-

plex under investigation. UV-laser footprinting eliminates several of these

disadvantages, but also creates others (see Subheading 1.4.).

1.3. Principle of UV-Laser Footprinting

The principle of ultraviolet (UV)-laser footprinting is schematized in Fig. 1.

The sample containing the nucleoprotein complex is irradiated with a short

(less than 10 ns) pulse of UV-laser light. An identical sample, but lacking the

162 Geiselmann and Boccard

protein, is treated in parallel in the same way. The conditions are adjusted such

that the number of photons delivered onto the sample exceeds the number of

absorbing molecules. The nucleic acid bases are excited and undergo photo-

reactions, the nature of which depend exquisitely on the local environment of

the bases. The possible reactions include intrastrand reactions of consecutive

bases (the formation of thymine dimers is the most prominent such reaction),

interstrand reactions (although their quantum efficiency is too low to contrib-

Fig. 1. The principle of UV-laser footprinting. The double-stranded DNA is repre-

sented by the double line, and the protein by the ellipse. (1) A sample of DNA alone,

or the DNA–protein complex is irradiated with one pulse of UV-laser light. The bases

can undergo intramolecular photoreactions, react with the solvent (symbolized by the

circle), or form a crosslink with the bound protein (symbolized by the line connecting

DNA and protein). Only the reactions of the top strand are shown in the figure. (2) The

samples are denatured, a radioactively labeled primer (line with diamond) is annealed

to one stand of the DNA (the top strand) and extended using a DNA polymerase (dot-

ted line). The primer extension stops at damaged bases or at the end of the fragment.

(3) The primer extension products are analyzed on a sequencing gel, along with a

dideoxy-sequencing reaction using the same primer. The location of the photoreac-

tions are marked with an arrow, the crosslink and the runoff are indicated.

UV-Laser Footprinting 163

ute to footprinting signals), reactions with solvent molecules (e.g., with H

O),

and crosslinks with the protein (4–8).

In general, it is not possible, but neither is it necessary, to determine the

nature of the photoreaction at a particular base. Because the photoreactions are

highly sensitive to the local environment of the DNA, binding of a protein

changes this environment and thus produces a footprint (i.e., a difference

between the DNA photoreactivity in the absence and presence of the protein).

In the example of Fig. 1, the photoreaction at the left extremity of the DNA

fragment remains constant because the protein does not change the local envi-

ronment of this base. However, the presence of the protein prevents a photo-

reaction toward the right end of the fragment (perhaps by excluding water from

the vicinity of the base) and favors a new photoreaction with an amino acid

side chain, resulting in a covalent bond, a crosslink, between the protein and

the DNA. All such reactions modify the nucleotide base and, hence, impede

the progression of DNA polymerase during subsequent replication of the dam-

aged DNA strand. Arrest of the polymerase is detected in a primer extension

reaction by the appearance of shortened replication products, which then serve

to localize the modified base.

1.4. Advantages and Disadvantages of UV-Laser Footprinting

Ultraviolet-laser footprinting circumvents many potential artifacts of more

classical techniques by trapping the complex under investigation at the time of

irradiation. The signal is acquired very rapidly (on the order of microseconds)

(i.e., faster than typical rearrangements of a DNA–protein complex [on the

order of milliseconds]), and the footprint thereby freezes the initial state of the

complex (9). Laser light, as opposed to ordinary UV light, is needed in order to

limit the “incubation time” to several ns while providing a sufficient number of

photons to excite all nucleotide bases of the sample. The rapidity of signal

acquisition allows one to obtain kinetic structural signals by irradiating a com-

plex at different times after mixing the components. The technique has the

further advantage that it can be transposed to in vivo experiments in a straight-

forward manner because UV light readily penetrates bacterial cells. However,

because of the limited sample size of in vivo experiments, the present version

of the technique is only suitable for studying DNA–protein interactions in bac-

teria when using a binding site carried on a multicopy plasmid.

Using the UV-laser technology, it is possible to follow the kinetics of the

assembly of DNA–protein complexes. This technique can be used to determine

the order in which protein–DNA contacts are established in a multiprotein–DNA

complex. Proteins and DNA are mixed at time zero and the sample is irradiated

at a specific time interval after mixing (10). Modern mixing techniques allow a

time resolution on the order of a several milliseconds (11).

164 Geiselmann and Boccard

The main disadvantage of UV-laser footprinting is the unpredictability of

the footprinting signal. It should be noted that the footprint is not the result of

the protein shielding the DNA from the UV irradiation. The presence of the

protein merely changes the probability of certain photoreactions by excluding

solvent, changing the conformation of the DNA, or juxtaposing a reactive

amino acid and a particular base. A structural interpretation of the footprinting

signal is therefore, in general, not possible. A more practical limitation of the

technique is the need for a relatively expensive laser.

We describe the experiment for the particular case of the binding of an

Escherichia coli protein, the integration host factor (IHF), to one of its specific

binding sites, the yjbE site (12), in vitro and in vivo. IHF is a small,

heterodimeric protein (molecular weight [MW] of the dimer is approx 19 kDa)

that binds to specific sites on the DNA (for a review, see ref. 13). Upon bind-

ing to DNA IHF bends the DNA by about 180º (14). This DNA bending gives

rise to a very strong UV-laser footprinting signal, probably because two con-

secutive pyrimidines are brought into optimal alignment for the formation of a

pyrimidine dimer (15). Variations of this basic protocol, applicable to the study

of any DNA–protein interaction, are described in Subheading 4.

2. Materials

1. Ultraviolet-laser: The most commonly used lasers are YAG lasers, e.g., the Spec-

tra-Physics Quanta-Ray GCR series lasers, which emit infrared light of 1064 nm.

Two consecutive passes through a frequency-doubling crystal yields high-inten-

sity light of 266 nm, which is sufficiently close to the absorption maximum of

nucleic acid bases. Frequency doublers are a standard add-on for virtually all

commercially available YAG lasers. The laser power should be between 30 and

50 mJ per pulse at 266 nm and a pulse duration of 5 ns is standard. The energy of

one 30-mJ pulse represents about 4 × 10

photons (i.e., 67 nmol photons).

2. Power meter to measure the energy of the laser beam.

3. Thermostated water bath.

4. Spectrophotometer.

5. Water pump and 0.45-µM filter device for washing E. coli cells.

6. Phosphoimager: A phospho-storage device is ideal for quantifying sequencing

gels. The most commonly used instruments are sold by Molecular Dynamics,

Fuji, or by Bio-Rad.

7. IHF binding buffer: 50 mM Tris-HCl, pH 7.5, 70 mM KCl, 7 mM MgCl

, 3 mM

CaCl

, 1 mM EDTA, 10% glycerol, 200 mg/mL bovine serum albumin (BSA),

and 1 mM β-mercaptoethanol.

8. IHF: working stock solution at 3–5 µM in 50 mM Tris-HCl, pH 7.4, 800 mM KCl,

40 mM K-phosphate, 2 mg/mL BSA, and 10% glycerol.

9. Plasmid pBluescriptII (Strategene) containing the IHF binding site cloned into

the multicloning site (MCS). Stock solution in water at 100 nM.

UV-Laser Footprinting 165

10. Primer extension reaction. Annealing buffer: 1 M Tris-HCl, pH 7.6, 100 mM

MgCl

, and 160 mM dithiothreitol (DTT). Elongation mix for 2 µL: 2 U of T7

DNA polymerase in 2.4 mM of each deoxyribonucleotide, 3 mM Tris-HCl,

pH 7.5, 0.75 mM DTT, 15 mg/mL BSA, and 0.75% glycerol (see Subheading 3.

for the annealing and elongation steps). Stop solution: 95% formamide, 20 mM

EDTA, 0.05% bromophenol blue, and 0.05% xylene cyanol.

11. Sequencing reaction. Enzyme dilution buffer: 20 mM Tris-HCl, pH 7.5, 5 mM DTT,

0.1 mg/mL BSA, and 5% glycerol. Stop solution: 95% formamide, 1 mM EDTA,

0.05% bromophenol blue, 0.05% xylene cyanol.

12. Sequencing gel: Sequencing reactions are analyzed on 40-cm-long denaturing

(7 M urea) 8% polyacrylamide (ratio acrylamide:bis acrylamide 19:1) gels in 90 mM

Tris–borate, and 2 mM EDTA (TBE), and TBE used as running buffer. Gels were

transferred onto Whatman (3MM Chr) paper and dried at 80°C for 40 min in a gel

dryer (Bio-Rad model 583) linked to a vacuum pump.

13. Extraction of plasmid DNA. Solution I: 100 mM Tris-HCl, pH 7.5, 10 mM EDTA,

400 µg/mL RNase I. Solution II: 0.2 N NaOH, and 1% sodium dodecyl sulfate

(SDS) made freshly. Solution III: 3 M potassium, and 5 M acetate solution, made

by adding 11.5 mL of glacial acetic acid and 28.5 mL of H

O to 60 mL of 5 M

potassium acetate.

14. LB and minimal M9 media are used to grow and wash E. coli cells, respectively (16).

3. Methods

3.1. In Vitro UV-Laser Footprinting

1. Arrange the laser beam, using appropriate mirrors, such that it is directed verti-

cally into a water bath.

2. Align, and fix firmly, an Eppendorf holder in the water bath such that the laser

beam enters precisely in the center of an open Eppendorf tube. A piece of black

paper stuck into the bottom of the Eppendorf tube can help align the tube with the

laser beam; the impact of the laser light is very audible and “burns” the site of

impact, whitening the otherwise black paper.

3. Operate the laser in repetition mode (see Note 1) for at least 10 min and adjust the

doubling crystals to obtain a laser power at 266 nm of at least 30 mJ per pulse.

This laser power is measured during the warm-up period with an appropriate

power meter, before the actual footprinting reaction.

4. Incubate IHF at the desired concentration with 5 nM plasmid DNA in 40 µL

binding buffer for 20 min at 25°C (see Note 2). It is best to use a flat-bottomed

Eppendorf tube, but regular 1.5-mL Eppendorf tubes are adequate. The laser beam

generally has a diameter of 5 mm, and for maximal use of the light energy, the

sample should have roughly the same dimensions. Care should be taken to ensure

that all of the sample is irradiated by the laser beam.

5. Place the sample under the laser beam and irradiate with one pulse of UV-laser

light. It is best to operate the laser in repetition mode (i.e., continuously emitting

around 10 pulses per second). Most lasers also have the possibility to emit a

single pulse of light. However, the power of such a pulse is not very well con-

166 Geiselmann and Boccard

trolled, because the yield of the doubling crystals is extremely sensitive to

temperature. Continuous emission of laser pulses at a frequency of about 10 Hz

ensures a constant temperature of the crystals and therefore a stable pulse energy.

An electronically controlled shutter is used to obstruct the beam. The shutter

opening is coordinated with the emission of the laser pulses to ensure that only

one pulse of laser light passes for each opening of the shutter.

6. After irradiation, remove the protein from the irradiated DNA by incubating with

50 µg/mL proteinase K for 15 min at 50°C. Extract the samples with half a vol-

ume of a phenol/chloroform solution (made by adding equal volumes of phenol

pH 8 and chloroform), precipitate with 2 vol of ethanol, and resuspend the DNA

in 18.5 µL of H

7. Primer extension (see Note 3): Add 2 µL of a solution of 0.2 µM radiolabeled

primer (5'-[

P] labeled using T4-kinase) to 18.5 mL of DNA. (Increasing the

primer concentration beyond this twofold excess over template will increase the

strength of the primer extension signals only marginally.) Denature the samples

by heating to 100°C for 3 min and chill on ice for 5 min. After the addition of 2.5 µL

of annealing buffer, incubate the samples for 3 min at 50°C (to anneal primer)

and chill again for 5 min on ice. Add 2 µL of the elongation mix (see Note 4) and

incubate the reaction for 10 min at 37°C. The mix is prepared freshly but can be

kept on ice for several hours.

8. Precipitate the DNA by adding 150 µL of ethanol and incubating for 10 min at

–20°C. Centrifuge for 10 min at full speed (about 12,000g) in a microcentrifuge.

Resuspend the DNA in 10 µL of loading dye.

9. Analyze 3 µL by gel electrophoresis on a denaturing 8% polyacrylamide sequenc-

ing gel. After electrophoresis, transfer the gel onto Whatman paper and vacuum

dry at 80°C for 40 min.

10. To determine precisely the location of the footprint, a reference ladder is gener-

ated by sequencing the same plasmid DNA using the same radiolabeled primer

(see Note 5). Denature 15 nM of plasmid DNA in a volume of 8 µL by heating to

100°C for 2 min. Chill on ice for 5 min. Add 1 µL of radiolabeled primer (0.25 µM)

and 1 µL of annealing buffer. Incubate for 3 min at 50°C. Chill annealing reac-

tion on ice for 5 min. Add 2.8 µL of each Deaza G/A T7 Sequencing™ Mixes

(Pharmacia) to four termination reaction tubes and prewarm at 37°C. Dilute 1 µL

of T7 DNA polymerase with 4 µL of enzyme dilution buffer. Add 2 µL of diluted

T7 DNA polymerase to the annealing reaction. Dispense 2.8 µL of annealing

reaction in each of the termination tubes. Incubate for 5 min at 37°C. Add 4 µL of

the stop solution.

11. Expose the dried gel to a phospoimager screen overnight and scan the screen

using the phospho-imager and its associated software (e.g., ImageQuant from

Molecular Dynamics).

12. Deduce a line profile of the different lanes using the Phospho-Imager software.

13. Transfer the data to Microsoft Excel and superimpose the scans on the same graph

in order to visualize and quantify the footprint (see Notes 6–11).

UV-Laser Footprinting 167

3.2. In Vivo UV-Laser Footprinting

1. Grow cultures of IHF

(W3110) and IHF

–

(W3110 hip) strains (12) transformed

with the plasmid carrying the ihf site in LB medium to the desired OD

600

(0.6 or

to saturation).

2. Wash the cells in minimal M9 medium (optically transparent buffer), resuspend

in minimal M9 medium to a final OD

600

of 1, and incubate the cells at 37°C (see

Note 12).

3. Irradiate as described above a large number (40–60), of 50-µL cell aliquots (see

Note 13). Freeze the cells immediately after irradiation in a dry-ice bath.

4. Pool the cells and extract plasmid DNA from 2–3 mL of cells using the following

alkaline lysis procedure. Centrifuge the bacteria for 5 min in a tabletop Eppendorf

centrifuge. Resuspend the pellet in 100 µL of solution I and add 100 µL of solu-

tion II. Mix by inversion several times. Add 100 µL of solution III and mix by

inverting the tube several times. Centrifuge the tubes at full speed in a tabletop

Eppendorf centrifuge (>10,000g) for 5 min. Transfer the supernatant to a clean

tube and add 210 µL of isopropanol. Incubate for 5 min at room temperature and

centrifuge the tubes at full speed for 10 min. Wash the pellet with 70% ethanol

and dissolved the DNA in 37 µL of H

5. For primer extension, add 2 µL of a solution of 0.2 µM primer (5'-[

P] labeled)

to 18.5 mL of DNA and process and analyze the samples in the same way as for

the in vitro reactions, steps 7–13 of Subheading 3.1. (see Notes 10 and 11).

4. Notes

1. Laser setup. As mentioned in the Subheading 3. in order to obtain a stable laser

power it is best to operate the laser in repetition mode. If a single-pulse mode is

used the energy of a particular pulse is ill-defined and the absence of a

footprinting signal may simply be the consequence of diminishing laser power.

All lasers provide an electrical signal that allows external equipment to be

coordinated with the laser pulse. To our knowledge, shutters are not commer-

cially available. However, it is an easy task for a good mechanics shop to con-

struct such a shutter. We used a shutter made of a small sheet of blackened Teflon

obstructing a hole of approx 8-mm in diameter through which the laser beam had

to pass in order to reach the sample. Any material can be used, but it should be

kept in mind that the laser will eventually burn a hole into the material and that it

is best to use a black material in order to minimize hazardous reflections of the

laser light. A simple electronic circuit controlled the opening of the shutter by

activating an electomagnet that pulled the Teflon sheet (via an attached piece of

metal) away from the hole. After the light had passed, the current to the magnet

was cut and a spring pulled the sheet back over the hole.

2. The binding buffer described is the standard buffer used for measuring DNA

binding of IHF. Other proteins may require different buffer conditions. The buffer

may be adjusted with certain limitations. The salt concentration should not be too

low in order to avoid nonspecific binding of the protein to DNA. The buffer

should not include a high concentration of reagents that absorb at 266 nm. For

168 Geiselmann and Boccard

example, the interaction of the cyclic AMP receptor protein (CRP) with DNA

requires the presence of cAMP in the buffer (17). Keep the concentration of such

nucleotides below 100 µM. As mentioned in Subheading 3., one laser pulse con-

tains the equivalent of about 100 nmol in photons. A typical reaction volume is

50 µL; therefore 100 µM ATP absorbs about 5 nmol of photons. A more physi-

ological concentration of ATP in the millimolar range would dramatically

decrease the yield of the photoreaction.

3. The procedure describes the footprinting reaction for only one strand of DNA.

Evidently, the other strand can be analyzed in the same way using the appropriate

primer. The primers should be chosen such that the region of interest is within

200 nucleotides from the primer. Most photoreactions have a quantum efficiency

below 1%, the formation of thymine dimers reaching several percent. Therefore,

on average, there will be roughly 1 photoreaction per 100 base pairs (bp).

Considering only the 100-bp region downstream of a primer assures single-hit

conditions. If the DNA carries too many photoreactions, the primer extension

will stop at the first defect and the signal of a photoreaction further downstream

will pass undetected.

4. The detection of photoreactions. All DNA polymerases are very sensitive to dam-

aged bases. It is not important to use T7 DNA polymerase in the primer extension

reaction, any other DNA polymerase (e.g., Klenow, Taq) gives equivalent sig-

nals. However, signals obtained with different polymerases may not necessarily

be identical because some may be more sensitive to particular photodamaged

bases than others. Even RNA polymerases can be used if the template harbors an

appropriate promoter. We have successfully used T7 RNA polymerase to tran-

scribe the region of interest from a T7 promoter located on the vector DNA. The

major signal of the IHF footprint remained unchanged, but, instead of a single

band, T7 RNA polymerase generates a doublet of bands (15).

5. To determine the location of termination sites precisely, we generated a refer-

ence DNA ladder consisting of a sequencing reaction of the same DNA region

(see Subheading 3.). In general, it is assumed that the primer extension reaction

of the irradiated DNA stops just before the modified base. For example, if the

sequence of the DNA read on the gel is 5'-GGAC-3' and the primer extension

reaction of the irradiated DNA shows a band at the position of the A in the

sequencing reaction (run in parallel on the gel), then the photoreaction most likely

took place on the following base pair (the C in the above sequence). Because the

primer extension reaction reads the opposite strand, the photoreaction actually

damaged the G marked with an asterisk:

5'-GGAC-3'

3'-CCTG

-5'

6. There are several possible reasons for not detecting a UV-laser footprinting sig-

nal. The most obvious problem is that the protein does not bind to the DNA.

Generally, we verify binding by a gel retardation assay. The second reason may

be that protein binding does not change the photoreactivity of the bases in the

recognition site. This is the case, for example, for CRP, which yields very weak

UV-Laser Footprinting 169

signals in UV-laser footprinting despite a strong interaction measured by other

techniques (J. Geiselmann, unpublished results). Because photoreactivity

depends on the sequence, a particular site might not be photosensitive; for

example, the main IHF signal had not been observed for the ssb site, probably

because the sequence of this site does not contain the highly reactive TC pyrimi-

dine doublet (15).

7. A trivial reason for not detecting a photoreaction is that the laser did not hit the

sample. The best control, and a control to include in all laser footprinting reac-

tions, is to analyze the primer extension reaction of DNA alone. Two DNA-alone

samples should always be included: one irradiated sample and an identical sample

that has not been exposed to UV-laser light. The irradiated sample should yield

readily visible bands all along the lane (Fig. 2, lane f), whereas the nonirradiated

sample should show no elongation arrests (Fig. 2, lane e).

8. The irradiated DNA does not give any elongation arrests with T7 DNA poly-

merase. Verify the primer extension mix. Perform a control primer extension

reaction using the nonirradiated plasmid template, but cut about 100 bp down-

stream of the primer with a convenient restriction enzyme. Primer extension using

this template should give a very strong band corresponding to the elongation

reaction reaching the end of the fragment.

9. Artifactual footprinting signals. A contaminated template preparation or a dam-

aged DNA template could be misinterpreted as giving a footprinting signal. It is

very important to verify that the nonirradiated template does not produce any

bands during the primer extension reaction. This is particularly important for in

vivo reactions because the plasmid preparation could partially damage the plas-

mid and lead to artifactual bands (Fig. 2, lanes c and d).

10. Comparing signals obtained under different conditions in vitro, or comparing in

vitro to in vivo signals. The efficiency of sample preparation or of the primer exten-

sion reaction can vary from sample to sample. In order to compare different lanes

we run a small portion (10%) of the primer extension reactions on a sequencing gel

and quantify the lanes using a phosphoimager. We then load the same samples on a

second sequencing gel, but equilibrating the amounts loaded according to the sig-

nal intensities on the first gel. For example, if one of the reactions was only half as

efficient as the other ones, we load two times more of this sample on the second

gel. This readjustment should not be allowed to exceed a factor of 2.

11. Quantifying lanes. Once an intensity equilibrated gel exposure is obtained, we

obtain line profiles of all lanes using the Phospo-Imager software and compare

lanes by superimposing the line profiles in Microsoft Excel. Remaining small

(several percent) differences in the intensities of the lanes should be normalized

by multiplying the scans with a scaling factor between 0.9 and 1.1 that is deter-

mined subjectively by the user in such a way that the global patterns superim-

pose. Lanes containing a high background of nonspecific radioactivity cannot be

quantified with confidence.

12. It is important to keep in vivo samples at the physiological temperature (37°C for

E. coli) to ensure optimal binding.