This protocol assumes that the user has already performed gel-shift experiments with their fragment and determined optimal binding conditions.

I.Preparation of DNA

The fragment of DNA used for footprinting should not be more than ~400 bp long; less is preferable. Make sure that your fragment has appropriate restriction sites: one should leave a 5í overhang (susceptible to Klenow fill-in), and the other should not. Gel shift experiments will have given an indication of what sequences are bound by protein; it is desirable to create a version of the fragment in which hypothesized binding sequences are mutated. For easy comparison, mutant and wild-type versions of the fragment should be the same length when digested.

1. Grow up 100 mL of bacteria containing your plasmid(s). Purify by Qiagen midi-prep according to manufacturerís directions. Quantitate the resultant DNA.

2. Digesting DNA: In my case, I had 1-1.5 ug of DNA per ul from the Qiagen prep.

    A. In order to get lots of DNA, set up a total of 20 digests, of 100 ul each, with 1-2 ul DNA and 3 ul of each restriction enzyme.  Allow to digest a full 24 hours.

    B.Combine digests into 5 tubes of 400 ul each; phenol-chloroform extract and ethanol-precipitate.

    C. Resuspend each pellet into 20 ul TE.  Run on a 2-3% agarose gel.  I used a small (10 cm) gel and a 6-well comb.  After running it out for an appropriate period of time, purify the DNA.  I did so by excising the strip containing the DNA and electroeluting according to Sambrook et al (with the exception that I reversed the current for only 30" since I was working with fairly short fragments).

    D. Resuspend DNA in 25-50 ul TE.  Quantitate by running a small amount into an agarose gel against known standards.  You should end up with 20-30 ug.

3. Labeling DNA: Assuming that two pieces of DNA are to be labeled (one wild-type, one mutant), order 1 mCi of radiolabeled dNTP about a week in advance of labeling; in my experience, it took that long to arrive. Once the radiolabeled dNTP is added, these reactions are VERY HOT.  Work behind shields (I used Plexiglass plus lead foil) and dispose of radioactive material appropriately.

    A. Combine
            65 ul DNA + TE
            50 ul 32P-dATP (or labeled dNTP of choice)
            60 ul 5 mM cold dCTP, dGTP, dTTP (or appropriate cold dNTPs)
            10 ul NEB2
            5 ul Klenow fragment

    B. Vortex briefly (and carefully!), and spin down.  Incubate behind shields at room temperature for 20í.

    C. Do a cold chase by adding 60 ul 5 mM dATP (or appropriate cold dNTP).  Vortex and spin down.  Incubate at room temperature 5í

    D. Phenol-chloroform extract and ethanol-precipitate the DNA.

    E. Resuspend in 50 ul ddH2O.  Perform scintillation counting on a small amount of DNA (0.5 ?1 ul).  I typically got 45-80 million cpm per labelling reaction.
 

II.Calibration of DNAseI reactions

 In a footprinting experiment, one wishes to obtain a nice, even ladder of fragments in which the intensity of smaller fragments is close to the intensity of larger fragments after DNAseI digestion.  Because every piece of DNA differs in its susceptibility to DNAseI, it is necessary to determine digestion conditions experimentally.  Wild-type and mutant versions of a particular fragment may differ slightly.

Digestion reactions will contain the following (recipes below):
    radiolabeled DNA
    binding buffer (as for gel-shift experiments) at appropriate concentration
    other chemicals used in gel-shifts (e.g. KCl, DTT)
    ddH2O
    DNAseI (Boehringer-Mannheim)
    digestion buffer

    I recommend ~1 million cpm of DNA per reaction to start.  For a 100-150 bp fragment, this is a very suitable amount for a short exposure at room temperature (2-3 hrs).  For a longer fragment, a longer exposure may be required (i.e. overnight).
    Remember that a 100-bp fragment will produce a higher-intensity ladder than a 200-bp fragment if the same number of cpms/reaction are used.  Why? Digestion of a 200-bp fragment results in a greater number of digestion products (= a longer ladder) than digestion of a 100-bp fragment, and so less label is associated with any digestion product from the 200-bp fragment than the 100-bp fragment.  During the process of DNAseI calibration, play with cpm/reaction and gel exposure time to obtain a satisfactory ladder intensity.

Recipes:

5X basic-Tris DNA-protein binding buffer (1 mL): (or use your own DNA-protein binding buffer)
    50 ul 2 M Tris
    10 ul 500 mM EDTA
    25 ul 1 M MgCl2
    10 ul 10 mg/ml BSA
    5 ul Igepal
    500 ul 100% glycerol
    400 ul ddH2O

Digestion buffer (1 mL):
     400 ul 1 M HEPES pH 7.5
     150 ul 1 M MgCl2
     100 ul 500 mM CaCl2
     350 ul ddH2O

Quench solution (10 mL):
    2 mL 0.5 M EDTA pH 5.0
    0.6 mL 1 M Tris pH 8.0
    7.4 mL ddH2O

Hereís an initial set of conditions to try in the calibration of DNAseI with your fragment.  Dilute DNAseI in its buffer.  Donít forget to include an undigested control!
    1 ul 1/10 dilution DNAseI vortexed for 15", 30", 45", and 1í
    1 ul 1/20 dilution DNAseI vortexed for 15", 30", 45", and 1í
    1 ul 1/100 dilution DNAseI vortexed for 15", 30", 45", and 1í

    Before starting, dilute sufficient DNA to a concentration of 1 million cpm/ul.  Label reaction tubes and write down which tube number corresponds to which set of reaction conditions.
    Including DNAseI and digestion buffer, the reaction mix will total 100 ul.

1. Mix:
        ~1 million cpm DNA (1 ul of diluted DNA)
        Binding buffer to appropriate concentration (I suggest starting with a 5x solution)
        Other chemicals required for good protein-DNA interaction as determined by gel-shift
        ddH2O to 90 ul (including DNAseI to be added)

2. After addition of DNA, perform scintillation counting on reaction tubes.  Add more DNA as needed to even out the cpms in each reaction. Differences of no more than ~100,000 cpm between reactions are acceptable.

3. Vortex and spin down tubes.

4. Add 10 ul digestion buffer to one tube.
    Immediately add the appropriate amount of DNAseI of the appropriate concentration (or add nothing if this tube is the undigested control), and vortex well for the appropriate amount of time.
    Immediately add 100 ul of quench solution to the tube and vortex well for a few seconds.  Place tube on ice.
    Repeat until all reactions are complete and on ice.

5. Phenol-chloroform extract all reactions and spin at 14K rpm for 15í.  Remove aqueous layer - DO NOT TAKE UP ANY OF THE INTERPHASE.  In my hands, this prevented overly salty reactions at loading time

6. Ethanol-precipitate by adding 1 ul 5 mg/ml glycogen, 20 ul 3M NaOAc, and 400 ul 100% EtOH.
    Carefully invert tubes a few times and incubate at room temperature 10í.
    Spin at 14K rpm for 15í.  The pellet should be pretty small, about the size of a DNA pellet resulting from a Qiagen Miniprep.  If the pellet is large, a great deal of salt is present.  The gel will run very strangely as a result.  (See Troubleshooting for more on the salt problem.)
    Rinse with 500 ul 70% EtOH and spin at 14K rpm 5í.
    Carefully remove all liquid with a long filter-tip, leave the tubes open, and air-dry for 10í.

7. Resuspend DNA in 5 ul ddH2O.  Incubate at 37° for 10-60í.
    Add 5 ul 2x formamide loading buffer (as for sequencing reactions; see Sambrook et al. for recipe) and mix.  Reactions may be stored at ?20° overnight.
    Load 4 ul of each reaction onto a polyacrylamide sequencing gel.  Use a square-tooth comb with teeth ~4 mm wide and 1 mm width gaps between teeth.  In order to ensure polymerization, place 2-3 clamps in the center of the comb and an additional 2 clamps at each end.
    Run gel as for a standard sequencing gel.  Bromophenol blue runs at about 65 nt in an 8% sequencing gel, and that the labeled end of the DNA will run at the bottom of the gel.
    After running the gel, dry it and expose it to film for ~3 hours.  If this exposure time is insufficient, make a second longer exposure.

Common results of a DNAseI calibration gel:

• Assuming reactions were loaded in order from less-digested to more-digested, the gel should show a series of digestions in which lanes to the left are darker near the top (an excess of larger fragments, indicating underdigestion) and lanes to the right are darker near the bottom (an excess of smaller fragments indicating overdigestion).  Use the reaction conditions that result in a digestion ladder that is of even intensity over its entire length for all future footprinting experiments using that fragment.

• Sometimes even the smallest amount of DNAse tested for the shortest amount of time will overdigest a fragment of DNA.  If this occurs, retest using a smaller amount of DNAse and/or incubating for as short a period of time as possible (1-2 seconds).  You may wish to pre-load a pipet with 100 ul quenching solution and place it carefully to the side, so as to stop the reaction as quickly as possible.  Conversely, if all reaction conditions tested result in underdigestion, increase the amount of DNAseI used and/or the amount of time for digestion.

• If your gel ëfrownsí badly, there was an excess of salt in your reactions.  (See Troubleshooting, last page.)

• If nothing is visible on the gel after an overnight exposure, the pellet was lost or did not precipitate in the first place.  Try precipitating digested DNA for an hour at ?20° instead of at room temperature for 10í and proceed as above.

    1. Set up reactions using the optimal conditions determined by calibration.

    2. Before adding digestion buffer, add appropriate amounts of protein to reactions.  Include two controls: one in which neither protein nor DNAseI is added, and one in which no protein is added but DNAseI is.  Incubate the reactions at whatever time and temperature were used for gel-shift reactions.

    3. Spin down reactions, then add digestion buffer and proceed as above.

Be particularly careful about taking only the aqueous layer after phenol-chloroform extraction, since the interphase will have more garbage in it now than it did during calibration.
 

My reaction conditions follow:

~1 million cpm DNA (1 ul)
30 ul cocktail, which included:
    20 ul 5x binding buffer
    5 ul 1 mM DTT
    5 ul 50 mM KCl
Protein as appropriate (1-5 ul of various dilutions)
q.v with ddH2O to 90 ul (including DNAseI to be added)
 
    1. Mix all but protein together.  Add protein and vortex gently.  Spin down tube contents and incubate at 30°C for 20í.

    2. After incubation with protein, spin down tube contents.  Rapidly add 10 ul digestion buffer, add DNAseI as appropriate (1 or 3 ul of 1/100 dilution, depending on the fragment I was using), and vortex as short a time as possible. Immediately add 100 ul of quench solution, vortex well, and place tube on ice until all digestions have been completed.

    3. Add 200 ul phenol/chloroform, vortex well, and spin 15í at 14,000 rpm.  Take off top layer, being careful to take none of the interphase layer.

    4. To precipitate DNA, add 20 ul 1.5 M NaOAc, 1 ul 5 mg/ml glycogen, and 400 ul 100% EtOH.  Carefully invert tubes 1-2 times and incubate at room temperature for 10í.
Spin at 14K rpm 15í at room temperature.  Remove supernatant.

    5. Add 500 ul 70% EtOH,  and spin 5í at 14K rpm to rinse DNA.  Remove liquid, then add 500 ul 50% EtOH.  Spin 14K rpm for 5í to rinse DNA.  Remove all liquid and air-dry tubes for 10í.

    Resuspend pellets in 5 ul ddH2O.  Incubate at 37°C for 10-60í.  Add 5 ul of 2X formamide loading buffer.  Store overnight at ­20°C overnight before running on a sequencing gel.  Run with G+A ladders (treat with dimethyl sulfate and piperidine according to Sambrook et al).  Dry gel and expose to film 3 ­ 16 hrs.
 

IV. Troubleshooting: Too much salt!

This was the major problem that I encountered.  My initial gels ëfrownedí badly, and my lanes pinched together near the bottom of the gel.    I took several steps to alleviate this problem:
 

• not taking too much of the aqueous layer off after phenol-chloroform extraction.  I believe that the junk at and near the interphase pulled down salt during precipitation.

• adding NaOAc to only 0.15 M rather than 0.3 M for precipitation.  To determine this concentration, I split some digestions after phenol-chloroform extraction and added different amounts of NaOAc to each set of tubes.  In my hands, 0.15M NaOAc was a good concentration.  I also tried omitting glycogen from the precipitation reaction, but found that in my hands this resulted in a severe decrease in ladder intensity.

• performing precipitation at room temperature for 10í rather than at ­20° or ­80° as was originally suggested to me.

• performing two washes after precipitation: the first, 500 ul of 70% ethanol; the second, 500 ul of 50% ethanol.  I was nervous about using so low a concentration of ethanol for the second wash, but it helped eliminate salt without removing significant amounts of DNA

• loading an extra lane or two at each end.  I observed that the gel bottom tended to pinch less if there were more lanes loaded, and so I loaded two marker lanes on each side instead of one.  The outermost marker lanes were cropped in Photoshop for publication.