Protein-protein interactions

[Song Tan]

In article <smrodems-090394234355 at f181-102.net.wisc.edu>
smrodems at students.wisc.edu (Steve Rodems) writes:
>I've been doing gel shift experiments to analyze protein-DNA interactions
>and I've been using a Tris-glycing buffer system.  It is my understanding
>that tris-glycine is of higher ionic strength than other commonly used gel
>shift buffers (eg., TAE, TE, 0.5x TBE) and that with this system you tend
>NOT to detect weaker interactions.  My question is:  Using tris-glycine are
>protein-protein interactions stable?  That is, say a DNA binding protein
>binds DNA as a homodimer but another non-DNA binding protein can interact
>with the DNA binding protein through protein-protein interactions.  Can one
>detect the non-DNA binding protein as part of the complex using a
>tris-glycing buffer system.
>
>Does anyone have ideas as to what gel system and/or binding buffers (salt
>conc's) needed to detect protein-protein interactions as well as
>protein-DNA interactions?
>
>-- 

I'm not sure if it's really true that weaker interactions are more difficult to
detect with Tris-glycine buffer systems.  One point worth keeping in mind is
that glycine is a zwitterion and doesn't have any net charge at neutral pH's. 
To be sure, the glycine in Tris-glycine does contribute some charge at pH 8.3
or whatever the actual pH of Tris-glycine is.  But it's much less than adding
100 or 200 mM NaCl.

You can definitely use buffering systems such as TBE, 0.5xTBE, TAE and
Tris-glycine to detect the protein-protein interactions that you mentioned. 
I've worked on (still working on, actually) entirely analagous situations to
the one you described:  protein A binds to DNA, protein B doesn't bind to DNA
on its own, but does to the protein A/DNA complex.  In my case protein A is the
yeast transcription factor MCM1 and protein B is MATalpha1.  I use a different
formulation of TAE which I called EAT for my bandshifts (1xEAT is 3.3 mM NaAc,
6.7 mM Tris, 1 mM EDTA adjusted to pH 7.5 with HAc), but I have also used
0.5xTBE to form both the binary (MCM1/DNA) and ternary (alpha1/MCM1/DNA)
complexes.  I've stuck with EAT because this gel system provides better
resolution of the binary from the ternary complex compared to 0.5xTBE, probably
because of the lower pH.  Don't think I've ever tried the Tris-glycine system
for this experiment, I'm afraid.  The EAT mixture has _very_ low buffering
capacity as you might expect, but this is not a problem as long as you
recirculate the buffer.  [Reference for this work:  Tan, Ammerer and Richmond,
EMBO Journal, 1988, 7:4255-4264]

I'm pretty sure that I've seen Tris-glycine being used for situations similar
to what you described, although I don't have any references off hand.

I think the empirical approach of simply trying out different buffering systems
is the best.  I tend to try 0.5xTBE, EAT and Tris-glycine for bandshift
experiments.  At least one usually works, but if necessary one can always try
more exotic concoctions.  As to the binding buffers, 50 to 100 mM salt (NaCl,
KCl, KAc etc) seems to be most common but you will want to perform a salt
titration to find what's best for your situation.


Song Tan
Institute for Molecular Biology and Biophysics
ETH-Honggerberg (Swiss Federal Institute of Technology)
8093 Zurich, Switzerland
email:  tan at aeolus.vmsmail.ethz.ch