DNA-binding protein?

Tom Anderson ucgatan at ucl.ac.uk
Fri Aug 20 06:13:56 EST 2004

On Thu, 19 Aug 2004, Mauro Bongrazio wrote:

> I'm looking for the function of a new small protein (about 30 kD) which
> is supposed to have a nuclear localization.

Your saying 'supposed to' makes it sounds like you're not certain; if
that's the case, a good place to start might be immunofluorescence against
your protein, or expression of a GFP fusion, to see where it lives. Or
fractionation followed by SDS-PAGE and western blotting, but that's more
of a pain and less trustworthy, IMHO.

Note that proteins smaller than ~40 kDa can enter the nucleus by passive
diffusion, as i'm sure you're aware, so you'd expect to find a certain
amount of the protein in the nucleus - the question, i suppose, is whether
there's more than you'd expect through diffusion. You might try expressing
plain GFP (or RFP etc) as well, on the assumption that it will partition
equally between the cytoplasm and the nucleus (it's ~30 kDa), then doing a
ratiometric analysis to see if your protein partitions differentially.

> I believe it could bind DNA and have a role in the regulation of
> transcription. How can I investigate this hypothesis? Could you please
> help me to choose the right approach?

I haven't done work along these lines myself, but some colleagues have,
and the things they tried (that i know of) were:

- ChIP cloning, to identify binding sites
- Overexpression of the protein followed by western blotting for histone
  modifications (acetlyation, methylation, etc) to see if it is
  recruiting coactivators

Other things that spring to mind:

- Overexpression of or RNAi against the protein followed by microarray (or
  differential display, SAGE, etc) analysis to characterise the effects on
- A yeast 2-hybrid or coimmunoprecipitation screen to see if it interacts
  with known components of the transcriptional machinery
- Once you have candidate binding sites, gel shift assays to validate them
- RNAi your protein, then see if transfection of a mutant lacking the NLS
  will rescue; this requires that you see a phenotype when you RNAi, that
  you know where the NLS is, and that overexpression of the mutant doesn't
  generate a phenotype of its own
- Footprinting and crosslinking techniques; i'm not sure how to use these

Hope this helps!


Tom Anderson, MRC Laboratory for Molecular Cell Biology, UCL

More information about the Methods mailing list