ABSTRACT
For the analysis of protein-DNA interactions by coupled gel-shift/footprinting, DNA fragments need to be extracted from polyacrylamide gels and subsequently separated on high resolution gels. Due to impurities in the
extracted DNA, single nucleotide resolution is frequently not achieved. We now
describe an improved experimental strategy that employs transient coupling of
DNA fragments to a solid support in order to extract DNA of high purity
quantitatively, rapidly and reliably. As an example, we describe the
application of our protocol to the `in-gel footprinting' by copper phenanthroline. The method should also find
application to the chemical interference assays.
Sequence-specific protein-DNA interactions are conveniently analysed by either
electromobility shift assays (EMSA) or `footprinting' assays. The former
technique detects protein-DNA complexes through their slower migration as compared with free DNA
during electrophoresis in a native gel. It is able to detect protein binding
under conditions of DNA excess and yields a rough estimate of the mass and/or
conformation of the protein-DNA complexes. The latter technique is based on the fact that tight DNA-binding proteins protect their target sequences from nuclease digestion or chemical modification. It requires that all
DNA molecules in a reaction are bound by protein (i.e. protein excess) and enables
the mapping of binding sites with single nucleotide resolution. Protocols that
combine EMSA and the footprinting assay yield maximal information and are the
most suitable for analysis of complex protein mixtures. In the simplest case, protein-DNA complexes are treated with nucleases (or chemicals) before separation on a native gels. Nicked DNA from individual complexes
is purified and then analysed on high resolution sequencing gels. This protocol
will yield footprint data also if the target protein is only present in
subsaturating amounts, since homogeneous material is recovered from each band
in the native gel. In a modification of this assay, protein-DNA complexes are first separated from each other and from free DNA on a
native gel and the footprinting reaction is then performed in the gel (
1
).
The crucial step common to these protocols, as well as to the related
`interference' assays (
2
), is the isolation of the nicked or modified target DNA from the native gel,
which is sufficiently pure for optimal single nucleotide resolution on
sequencing gels. Trivial as it may seem, the failure to obtain DNA of a high
enough purity is frequently the reason for unsatisfactory results. Current
procedures-such as multiple successive organic extractions and precipitations, the
transfer of DNA onto NA-45 membrane by blotting, or electrophoresis followed by high salt elution-suffer from low recoveries, tricky hands-on steps or variability.
We describe a method for rapid and efficient purification of DNA from bandshift
gels yielding DNA of the highest purity for sequence gel analysis. The
strategy, which avoids organic extractions and precipitations, is derived from
`solid phase footprinting' (
3
). A DNA fragment, labelled with
32
P on one 5'-end and carrying a biotin on the other, is used for the EMSA. The
region of the gel containing the DNA of interest is excised from the gel. The
DNA is then eluted in the presence of proteinase K and immediately immobilised
onto streptavidin-coated paramagnetic beads therefore selecting for double-stranded DNA (biotin and
32
P label). Successive washes of the immobilised DNA yield a sample with unmatched purity. Denaturation upon addition of
formamide loading buffer to the DNA beads and release the labelled DNA
fragments for separation on a sequencing gel.
The method is applied to `in-gel footprinting' by copper phenanthroline (OP-Cu) (
1
). Chemical cleavage of DNA by OP-Cu is achieved by the binding of the 1,10 phenanthroline cuprous complex to
the minor groove of the DNA. Oxidation of the complex creates a highly reactive
intermediate that initiates the cleavage of the phosphodiester backbone in its vicinity. Protein-DNA interactions, as well as local distortions of the DNA structure, can be
visualised as protections from chemical cleavage or modulations of reactivity
respectively (
1
). The small size of the attacking chemicals improves the resolution of the
footprint, when compared to the use of bulky enzymatic nucleases, which are
sterically hindered to cleave close to the bound protein. The fact that the DNA
cleavage reagents diffuse uniformly, efficiently and rapidly through a gel
matrix leads to in-gel applications (
1
).
A 168 bp fragment containing HSE 1 and HSE 2 of the hsp70 promoter (
4
) was generated by PCR using one biotinylated and one radioactively labelled primer (
3
). The fragment was gel-purified and concentrated by successive isobutanol extractions. Standard binding
reaction (
3
) containing ~10
6
c.p.m. target fragment, 1 [mu]g poly(dI[middot]dC), 2 [mu]g BSA and increasing amounts of recombinant
Drosophila
HSF was set up and the resulting protein-DNA complexes were separated on a 1 mm thick 4% polyacrylamide gel in 0.5* TBE. The gel was subjected to Op-Cu treatment at room temperature exactly as described (
1
). Briefly, after the electrophoresis, the gel was immersed in a glass tray
containing 200 ml 10 mM Tris-HCl, pH 8. The OP-Cu complex solution (obtained by mixing 1 ml 40 mM ethanolic
solution of 1,10-phenanthroline monohydrate, 1 ml 9 mM CuSO
4
and 18 ml water) was added and the gel was slowly agitated for 5 min to allow
even diffusion of the chemical throughout the gel. The cleaving reaction was initiated by adding 20 ml 58 mM mercaptopropionic acid and
mixed on a shaking platform for 20 min. The reaction was stopped by adding 20 ml 30 mM neocuprine hydrate in ethanol.
The gel was rinsed three times with distilled water, wrapped with plastic foil
and autoradiographed for 15 min. Titration of HSF into the binding reaction
yields a pattern of three retarded bands (lst-3rd shift, Fig.
1
A). The areas corresponding to free DNA and to the 2nd and 3rd shifts were cut
from the gel and the intact gel pieces were incubated in 500 [mu]l TE (10 mM Tris-HCl pH 8.0, 1 mM EDTA) and 100 ng/[mu]l proteinase K overnight at 37oC. The liquid was transferred into a new tube, 500 [mu]l 5 M NaCl and 50 [mu]l (500 [mu]g) Dynabeads Streptavidin M-280 (Dynal, Oslo; washed according to the
manufacturer's recommendation) was added. The immobilisation occurred during a
30 min incubation at room temperature with occasional gentle agitation. The
liquid trapped in the lid was collected by a short spin and DNA beads were concentrated for 30 s on a magnetic particle concentrator (MPC, Dynal, Oslo). The supernatant was removed and the
beads were washed three times with TE. Finally the beads were suspended in 5 [mu]l formamide loading buffer, heated to 75oC and the released DNA fragments separated on a sequencing gel as
described (
3
). We routinely recover >60% of the radioactivity from the gel slice for separation on the sequencing gel. Figure
1
B demonstrates that the footprint obtained from the 2nd shift is due to binding
of HSF at HSE 1 (lane 2), whereas the more retarded 3rd shift contains HSF
bound to HSE 1 and HSE 2 (lane 1) (
4
).
*To whom correspondence should be addressed. Tel: +49 6221 387 490; Fax: +49
6221 387 518; Email: quivy@embl-heidelberg.de
REFERENCES
Return