From: John Philo <jphilo@earthlink.net>
  To  : Arturo J. Morales <art@scripps.edu>
  Date: Mon, 4 Jan 1999 21:50:57 -0800

RE: Interesting problem...

Art,

You seem to be saying that you are calculating a MW based only on a
sedimentation coefficient, but that is not possible.  Are you actually
fitting g(s*) data to get the diffusion coefficient too, or using
SVEDBERG,
or ???

If your complex is tight enough, and the kinetics slow enough, that
there is
no significant dissociation during the velocity run, then in principle
you
can get out its diffusion coefficient as well as the sedimentation
coefficient, and therefore get the MW.  However, the dissociation
kinetics
may be too fast, and the boundary widths will reflect in part the
dissociation-association kinetics as well as the diffusion.  In that
situation the MW estimates from Gaussian fits to g(s*) and/or SVEDBERG
are
bogus.

Your vbar question does at first appear to be a circular 'chicken and
egg'
problem: if you know the vbar, you can get the MW, which tells you the
stoichiometry, from which you can calculate the vbar, etc...

The way out of this circle of logic is to use a self-consistency approach.
For each assumed stoichiometry there is a corresponding vbar.  If using
that
vbar gives you a MW that is consistent with the stoichiometry that you
assumed, then you have the correct answer.

In principle if you have a tight complex you can determine the molar
ratio
of protein to RNA in the complex by varying their mixing ratio, and
looking
at how much protein or RNA is left over (i.e. not in the complex) and
runs
with the sedimentation coefficient of the free material.  (Of course you
can
also do this with size exclusion chromatography).  This is another good
test
of the consistency of your conclusions about the stoichiometry.

'Hope this helps,

John Philo
Alliance Protein Laboratories

-----Original Message-----
From: Arturo J. Morales [art@scripps.edu]">mailto:art@scripps.edu]
Sent: Monday, January 04, 1999 4:36 PM
To: rasmb@alpha.bbri.org
Subject: Interesting problem...



Hello all,

I have an interesting problem and I hope that anyone out there with more
expertise than me in AU can provide some insight... this list has been
great so far at helping me learn how to use the XL-I mostly on my own
(since I wasnt the one going to the Beckman training in our lab :)

Ok, here's the problem:

I have a protein-RNA interaction that I'm trying to quantify... I know it
binds... Kd in the low nm range...

If I do a velocity run on the RNA, I get about 3S which assuming vbar of
.53 gives me MW of ~25Kd (just about right for tRNA)... so far so good...

If I do the same on the protein, I get about 2S, which is about 24Kd
using
a vbar of .7595 (calculated from composition).  that works well too...
since it is expected to be a dimer (12kd monomers)

The problem arises when I mix them both... I end up getting about
2.5S-2.8S
(or so, I need to refine that part).  I know this is happening because
now
the vbar for the complex is somewhere in between .53 and .75, of course,
depending what vbar I use, I can get any MW I want... which is not really
that useful... My main question is, how can I approach this problem?  I
would like to get stoichoimetry (which I believe is 2:1 monomer:rna, but
it
could be 2:2...)  Is it acceptable to show the shift in S without
estimating MW? Can I determine vbar directly for a complex? what is the
best way?

I would appreciate any comments

THANKS!

Art

----------------------------------------------------------------------
     Arturo J. Morales    (RPI
'94)             art@scripps.edu
   Department of Biology - Massachusetts Institute of Technology &
  Skaggs Institute for Chemical Biology - Scripps Research Institute
                   http://schimmel.scripps.edu/~art