DNA Footprinting

Introduction

The various processes in transcriptional regulation have been studied extensively to determine the sequence specificity of DNA binding proteins.   One such molecular  technique is DNA footprinting, which can be defined as a method for assessing the selectivity of DNA sequence to a specific binding ligand. The DNase I and hydroxyl radicals are the most commonly used as footprinting probes in most experiments. Appropriate DNA substrate holds the key to the success of a footprinting experiment. DNA footprinting can be used for assessing transcription factors which bind to promoter, silencer or enhancer region of gene to regulate its expression.

Footprinting applications

The first footprinting methodology was developed in the year of 1978. The first method has its application in studying the interaction between DNA-binding proteins and their target sites. The above method was then used for assessing the sequence-specific interaction of ligands with DNA.  This adapted method also provides information about the sequence selectivity, affinity, and binding kinetics. This method was used to find the sequence specificity of various ligands such as actinomycin, echinomycin and related compounds, distamycin and other minor groove binders, mithramycin and many other ligands. This method is the go-to technique for assessing the sequence specificity of novel ligands. The method is applied to various types of ligands including polyamides, triplex-forming oligonucleotides, and minor groove binding ligands.

Principle

Footprinting can be simplified as a protection assay of inhibition of cleavage by ligand binding at the binding sites of a DNA fragment. The assay limits the action of  digestion by cleavage agent  such as DNase I or hydroxyl radicals. A double-stranded DNA is labeled at one end of each strand. Then this double-stranded DNA is cleaved by a chemical or enzymatic cleavage agent. A single-hit kinetic is adapted in the cleavage stage so that on average each DNA molecule is only cleaved once. In case of a cleavage agent which does not possess any sequence selectivity, randomly distributed products are obtained as end product.  The cleaved products are resolved using a denaturing polyacrylamide gel. Since the ligands are selective to a certain sequence, they will protect these regions from cleavage. These products will be missing from the reaction and will be found as a gap (‘‘footprint’’) in the gel.  The exact location of the binding sites for the ligand is determined by these gaps and they are found by running a control and ligand-treated digestion alongside suitable markers.  When the concentration of ligands varies over a range in a footprinting reaction, its binding affinity can be estimated. For assessing the reaction kinetics, in case of slow reactions the time dependence of the appearance of the footprint can be used as a parameter.

Choice of cleavage agent

 The characteristics of an ideal cleavage agent:

  • It should generate an even ladder of bands in ligand-free environment
  • Quick
  • Cheap
  • Easy to use
  • Should provide high-resolution information about the ligand’s preferred binding site.

Not a single agent known fulfills all these criteria. This leads to the use of a variety of enzymatic and chemical probes have been employed in different studies.

Choice of DNA substrate

The length of the substrates used for the footprinting are usually 50–200 base pairs long. The limiting factor for the length of the substrates is the resolution of the polyacrylamide gels. These substrates are restriction fragments, obtained from appropriate plasmids. The essential characteristic of substrates fragments should contain the preferred binding sites for the ligand under investigation. This attribute leads a problem when using highly selective ligands, especially if their selectivity is not studied prior to the experiment. There is a high possibility the possible dinucleotide will be represented in any given restriction fragment.  When the selectivity of ligands for only two consecutive basepairs are easily used for this technique. When ligand selectivity increases there is a chance of the preferred sequence is not included in the fragment.

Choice of label

Labeling radioactivity is still the preferred method of choice, fluorescence labeling also can be used. A fluorescent label is employed in case of automated DNA sequencing. The limitation of fluorescence labels is the sensitivity of the labels is not applicable in case of where target duplex system is in fentomol. Therefore radioactivity becomes the most used method of labeling.

The fragment can be labeled at either the 3′- or 5′-end. The source of the substrate and the location of the binding sites relative to the end of the fragment becomes the determinant factor for labeling. When Synthetic DNA oligonucleotides are used as footprinting substrates, labeling is carried out at the 5′-end. It is done using 32P-ATP and polynucleotide kinase. First, the complementary oligonucleotides are designed to produce single-stranded overhangs. These overhangs have the property to be labeled at the 3′-end by filling with Klenow fragment or reverse transcriptase.

The process of direct labeling of restriction fragments at the 5′-end is more laborious. This is due to the reason they possess a 5′-phosphate, which hinders the addition of 32P and therefore it has to be removed before proceeding to radiolabel. In cases of fragments produced by two restriction enzymes then the enzymes should be added sequentially. After the kinase reaction, only one strand is labeled in the target duplex 5′ end labeling can be efficiently done by PCR using 5′-end primers.  3′-end labeling is employed in case of restriction fragments. It is done by filling with  32P-dATP or 32P-dCTP using a suitable polymerase.

Methods

 Substrate preparation

5′-Labelling of synthetic oligonucleotides

  1. In this method, both strands of the synthetic duplex are synthesized, one of them is labeled and they are then annealed. One strand is first radiolabeled at the 5′-end using g-32P-ATP and polynucleotide kinase (PNK). g-32PATPis used at a concentration of 10 mCi/ml).
  2. The kinase reaction is usually very efficient and equimolar quantities of ATP and the oligonucleotide can be used (1 ml of 3 mM oligonucleotide, 1 ml g-32P-ATP, 2 ml PNK buffer, 1 ml PNK (10 units), 15 ml water).
  3. The reaction is complete after incubating at 37oC for 30–60 min.
  4. The labeled oligonucleotide can then be separated from any unincorporated label by gel filtration.
  5. Gel purify the labeled oligonucleotide, by adding 10 ml ‘‘DNase I stop solution’’ (10 mM EDTA, 1 mM NaOH, 0.1% bromophenol blue, 80% formamide) to the kinase reaction, boiling the reaction for 3 min before cooling on ice.
  6. Load the entire labeled nucleotide onto a denaturing polyacrylamide gel (40 cm long, 0.3 mm thick run at 1500 V for about 2 h).
  7. After electrophoresis, the glass plates are separated and the radioactive oligonucleotide is located by exposure to an X-ray film for about 1 min.
  8. Excise the band containing the radiolabelled DNA from the gel and the radioactive oligonucleotide is eluted by covering with (300 ml) 10 mM Tris–HCl, pH 7.5 containing 10 mM EDTA (TE) and gently agitating overnight at room temperature.
  9. The eluted DNA can then be concentrated by ethanol precipitation and redissolved in a suitable volume of TE buffer.
  10. Duplex DNA is then generated by adding an excess of the unlabelled complementary strand, heating to 100oC and cooling slowly to room temperature.
  11. The annealed duplex can then be further purified in a similar fashion using a non-denaturing polyacrylamide gel.

This method is suitable for generating radiolabelled duplexes with a synthetic oligonucleotide of 50 base pairs length. The problem with longer oligonucleotides is that the annealing reaction can generate some misannealed species, This is very often in cased of fragments containing several copies of related ligand binding sites. To counter this problem a different method of labeling is used.

  1. Synthesis only one strand of longer fragments.
  2. Amplifying this single strand by PCR, using appropriate primers, one of which has been 5′-end labeled with g-32P-ATP.
  3. Perform a few cycles with only the radiolabelled primer to ensure efficient incorporation of the radiolabel.
  4. follow this with several more cycles after adding more of this unlabelled primer.
  5. To avoid mis-annealing make the final step in the PCR reaction is an elongation step, not melting and annealing.

3′-Labelling

Cloning the footprinting template.

 3′-end labeled fragments are prepared by filling in the sticky-ends of appropriate restriction fragments with a-32P-dATP.

  1. In case of synthetic DNA fragments cloning them into the polylinker site of standard vectors is the easiest approach;
  2. Cloning of short oligonucleotide duplexes is done by annealing them with complementary synthetic oligonucleotides.
  3. PCR is preferred method for longer fragments. In the PCR method, full-length synthetic strand and appropriate primers are used to generate the full duplex.
  4. Remove the thermostable polymerase from the PCR reaction
  5. The amplified DNA is cut with BamHI.
  6. Purification these PCR products is done on 2% agarose gels.
  7. This is then ligated into BamHI-cut pUC18 and transformed into Escherichia coli TG2 and white colonies are selected from agar plates containing ampicillin, X-Gal, and IPTG.
  8. The sequence of the clones must be confirmed by sequencing.

Plasmid preparation

  • Since the radiolabelled DNA will be purified by gel electrophoresis, thereby removing any other impurities.
  • DNA plasmids purification can be done by cesium chloride density gradient centrifugation (in case if required).
  • The plasmid DNA is extracted from a 5 ml culture and this is eluted with 50 ml of elution buffer. This can be stored at -20oC for future use.

3′-End labeling

 The entire 50 ml plasmid stock prepared from a 5 ml culture is radiolabeled.

  1. To release the fragment the plasmid stock is first digested with two restriction enzymes. These enzymes should have the following properties:
  • They should be capable of cutting at opposite ends of the polylinker fragment and are typically HindIII–SacI, EcoRI–PstI, or HindIII– EcoRI.
  • At least one of these enzymes must have a 50-overhang containing a T (such as HindIII or EcoRI), whilst the other (such as SacI or PstI) should not and should preferably have a 30-overhang which cannot be filled in.
  1. After restriction digestion at 37oC for 1 h, 1 ml a-32P-dATP is added together with 0.5 ml AMV reverse transcriptase and incubated for 1 h at 37oC.
  2. The radiolabelled fragment is then separated from the remainder of the plasmid on a non-denaturing polyacrylamide gel.
  3. Twenty microlitres of 20% Ficoll solution containing 10 mM EDTA and 0.1% bromophenol blue is added to the radiolabelled mixture and the sample is applied to a 5–8% non-denaturing polyacrylamide gel.
  4. The gel (40 cm long, 0.3 mm thick) is run at 800 V in a 1*TBE running buffer for about 2 h, until the bromophenol blue has run most of the way down the gel.
  5. The glass plates are separated and the position of the labeled DNA fragment is established by short (1 min) exposure to an X-ray film.
  6. The relevant band is then cut from the gel and the radiolabeled DNA eluted by adding 300 ml TE buffer and gently agitating overnight at room temperature.
  7. The eluted DNA is finally precipitated with ethanol and resuspended in a suitable volume of TE buffer so as to give at least 10 counts per second/ml on a hand-held Geiger counter.

With fresh plasmid and a-32P-dATP, this process typically generates about 150 ml of radiolabelled footprinting fragment.

Footprinting

GA track marker

Location of purines in the the target DNA sequence is shown by a method called  GA marker track. This tep helps in identification of the sequences of the footprinting sites. This has often been achieved using Maxam–Gilbert chemical sequencing reactions with dimethylsulphate (specific for G) or formic acid (for G + A) followed by cleavage with hot piperidine. An alternate method follows: Mix 1.5 ml labelled DNA with 20 ml sterile water and 4 ml DNase I stop solution (10 mM EDTA, 1 mM NaOH, 0.1% bromophenol blue, 80% formamide). The sample is then incubated at 100 oC for about 40 min with the microcentrifuge tube cap open to allow evaporation.

DNase I footprinting

  1. 5 ml radiolabelled DNA is mixed with 1.5 ml ligand solution (usually dissolved in 10 mM Tris–HCl, pH 7.5 containing 10 mM NaCl for small molecules, though buffers such as 50 mM sodium acetate pH 5.0, supplemented with other ions are often required for experiments with DNA triplexes) and incubated at 20oC for at least 30 min.
  2. Longer incubation (overnight) is required for ligands which exhibit slow kinetics.
  3. Since equal volumes of ligand and DNA are mixed, the ligand concentrations should be prepared at twice the concentration required in the final mixture. For initial experiments, a wide range of ligand concentrations should be tested (say from 0.01–100 lM), though a narrower range can be employed once the approximate affinity has been estimated.
  4. After equilibration of the ligand–DNA complexes the mixture is digested by adding 2 ml DNase I, diluted in 20 mM NaCl, 2 mM MgCl2, 2 mM MnCl2. DNase I should have a stock concentration of 7200 U/ml.
  5. In order to achieve single hit kinetics about 90% of the total DNA should be undigested, though this level of cleavage is often exceeded, especially for longer DNA fragments.
  6. After 1 min the reaction is stopped by adding 4 ml DNase I stop solution (10 mM EDTA, 1 mM NaOH, 0.1% bromophenol blue, 80% formamide). There is no need to heat the samples at this stage and they can be stored for electrophoresis later.
  7. Before loading onto the gel the DNA is denatured by incubating at 100oC for 3 min and quickly cooled on ice before running on a denaturing polyacrylamide gel.
  8. Use an 8% gel for fragments of 150 base pairs and longer, 5 ml 10xTBE Buffer containing 8 M urea, 27 ml diluent (50% urea).
  9. Gels (40 cm long, 0.3 mm thick) are run at 1500 V for about 2 h until the dye reaches the bottom of the gel.
  10. The gel plates are then separated, the gel is fixed by immersing in 10% (v/v) acetic acid, transferred to Whatman 3MM paper and dried under vacuum at 80o
  11. The dried gel is then exposed to a phosphorimager screen overnight before scanning.

Hydroxyl radical footprinting

  1. For hydroxyl radical footprinting 2 ml radiolabelled DNA is mixed with 10 ml ligand solution (dissolved in 10 mM Tris–HCl, pH 7.5 containing 10 mM NaCl) and incubated at room temperature for 30 min. The excess volume of added ligand compared to DNA means that the ligand is hardly diluted on mixing. When the reaction is performed in smaller volumes that footprints are less clear.
  2. After 30 min incubation, the mixture is digested with 10 ml of a hydroxyl radical mix (as a fresh mixture of 400 mM ammonium ferrous sulphate, 2.5 lM EDTA, 10 mM ascorbic acid and 1% hydrogen peroxide mixed in the ratio 1:1:2:2).
  3. The reaction is incubated for 15–20 min before precipitating with ethanol in the usual way.
  4. Wash the precipitates several times (at least twice) with 70% ethanol (To get clear bands).
  5. The pellets are then resuspended in 8 ml DNase I stop solution (10 mM EDTA, 1 mM NaOH, 0.1% bromophenol blue, 80% formamide).
  6. The samples are then heated at 100oC for 3 min and rapidly cooled on ice before loading onto a denaturing polyacrylamide gel as described for DNase I footprinting.

Hydroxyl radical footprinting provides higher resolution footprints than DNase I footprinting, from which it is much easier to identify the exact ligand binding sites. However, hydroxyl radicals are not appropriate for quantitative footprinting studies. In addition, hydroxyl radicals often detect weaker, secondary binding sites that are not evident in DNase I footprints.

Association and dissociation kinetics

  1. A typical association reaction is initiated by mixing 15 ml of appropriately diluted ligand with 15 ml of radiolabelled DNA incubated at 20oC (or appropriate temperature).
  2. Note: It is important that under these conditions the ligand is present in a much higher concentration than its duplex target so that its concentration remains essentially unchanged throughout the reaction, which then approximates to pseudo-first order conditions.
  3. The reaction is followed by removing 3 ml aliquots from the mixture at various time points and digesting with 2 ml of DNase I (typically 0.03 U/ml). This is stopped after 20 s by the addition of 4 ll of DNase I stop solution.
  4. The concentration of DNase I is increased (about threefold) so as to minimize the time required for digestion, this allows for shorter time points. Since the binding reaction is proceeding during the digestion, short cleavage times are preferred. Time points are usually taken every 30 s, though with care it is possible to reduce the digestion to 10 s.
  5. Pseudo-first order rate constants are determined by fitting single exponential curves to the plots of band intensity (determined by densitometry or phosphorimaging) against time using a suitable graph fitting program. Bimolecular rate constants can be determined from the variation of the observed rate constant with ligand concentration. In some instances, these values can be determined from several bands within the target site.

Dissociation reactions

  1. Before initiating a typical dissociation reaction the ligand–DNA complex is formed by mixing 15 ml of appropriately diluted ligand with 15 ml of radiolabelled DNA, leaving the samples overnight at an appropriate temperature (e.g. 20oC).
  2. The reaction is then initiated by the addition of a suitable sequestering agent.It is necessary to demonstrate that the sequestering agent (excess of competitor DNA) does not affect the dissociation process; this can be demonstrated by showing that the measured dissociation rates are independent of the competitor
  3. The reaction is again followed by removing 3 ml aliquots at various time points (ranging between 30 s to overnight) and digesting each with DNase I for 20 s.
  4. For extremely slow dissociation reactions, such as those observed with DNA triplexes it may be necessary to increase the rate by working at higher temperatures.
  5. First order rate constants can be determined by fitting single exponential curves to the plots of band intensity (determined by densitometry or phosphorimaging) against time using a suitable graph fitting program.