More question about gel-shift assay

Jon Gilthorpe j-giltho at
Mon Jun 24 06:30:00 EST 1996

In Article <3DIJ99$xL1 at>, klebs.bbs at
(Leo) wrote:
> I am working on two different bacterial transcription factors.
> One is very easy to handle. I can get  beautiful gel-shift
> pictures easily.  However, I just can't get any discrete
> band out of the second transcription factor.  It always yielded
> a smear pattern. When you increase the  protein concentration,
> the DNA/protein complex just stuck in the wells. The protein
> is purified through His-Bind column and has been dialyzed.
> The probe is approximately 500 bp. Our typical buffer is
> 50 mM  HEPES pH7.8, 100 mM KCl and 2 mM MgCl2. We run the gel
> in 1 x TBE buffer at 4 degree C.  The binding
> , although smear, seems specific. Adding 50 x quantity of pBR322
> DNA in the reaction did not affect the smear pattern.
> I will be very grateful if any of you can provide me some thoughts
> to improve my experiment. Thank you.


As far as I can tell your problems stem from the size of your probe.  From
my experiences 500bp is really just too big even though you are working with
purified protein.  The shorter the probe the cleaner the shift especially
when working with ones >50bp. A maximal limit should be 250-300bp.

  I don't know if you just tried increasing the protein concentration, but
with large probes this will lead to everything shifting up into the well. 
You also need to increase the concentration of non specific competitor and
is best done by way of a titration. It is just a matter of balencing the
amount of protien/non specific competitor/probe, the latter of which must be
in excess.  

Having said all this, some bandshifts just dont look nice.  You could also
try to fiddle with the conditions.  If your factor is involved in
protein-protein interactions you can also try adding deoxycholate to the
reaction which dends to destabilise these but not protein-DNA ones.

Good luck.

Jon Gilthorpe
Division of Eukaryotic Molecular Genetics
National Institute for Mediacal Research

j-giltho at
(0181) 959 3666

More information about the Methods mailing list