[RASMB] Centrifugal force and protein dissociation

[John Philo]

Songpon,

With regard to the data analysis issues, you are being inconsistent about
whether you believe this is a reversible association or instead you are
looking at a mixture. When you fit multiple species to g(s*) distributions
you are assuming your sample is a mixture (and I note this issue is #2 on
the list of 'Caveats and Pitfalls' in the program Help file).

If you really had rapidly-reversible dissociation of your tetramer to dimer
and monomer then velocity experiments would not resolve the three
species---you would probably see only a single peak at some intermediate
sedimentation coefficient. For a reversible equilibrium you will only
resolve the individual species if the association-kinetics are very slow
compared to the rate of separation (i.e. hours). If the kinetics are indeed
that slow then in effect you do have a mixture and yes, the proportions of
the various species should reflect the equilibrium distribution that existed
at the initial loading concentration.

I suspect that your monomer and dimer are damaged species, not in
equilibrium with tetramer (or at least with much weaker association than
normal, undamaged subunits). As Art Rowe said, the key experiment you must
do is to change the protein concentration substantially (either in velocity
or sedimentation equilibrium). You also might consider running this sample
on size-exclusion chromatography, collecting the peaks, waiting a few hours,
and  then re-injecting them. If you have a mixture, each peak will run pure
the second time through, whereas if it is a slow equilibrium you will get
all 3 species from each collected peak.

With regard to your question about reliability of the fit, DCDT+ will
calculate the confidence intervals on the parameters, which is one guide. A
three-species fit with this approach is difficult because the resolution is
probably not good and because you must work with a very limited range of
data. DCDT+ has the ability to constrain the molecular weights to be
consistent with your monomer-dimer-tetramer assumption, and doing so should
make the remaining parameters more reliable.

Another useful approach to assessing reliability is to simulate your
experiment using a finite-element simulator and add realistic amounts of
random noise. If you can't reliably extract 3 species from synthetic data,
you certainly won't be able to do it with real experiments!

You would get more robust and reliable results using whole-boundary fitting
rather than the DCDT approach. With whole boundary analysis (available in
SVEDBERG, SEDFIT, LAMM, or ULTRASPIN) you get both better resolution of
species and higher signal/noise because you are using data covering
essentially the whole range of boundary movement. This also avoids the
problem of limited parameter accuracy from the DCDT approach that arises due
to certain approximations that must be made in that approach.

Having said all that, I should not forget to say that above all, the best
way to assess the reliability of your results is to REPEAT THE EXPERIMENT!

John Philo


-----Original Message-----
From: rasmb-admin at rasmb-email.bbri.org
[mailto:rasmb-admin at rasmb-email.bbri.org]On Behalf Of Songpon
Deechongkit
Sent: Tuesday, July 23, 2002 7:54 PM
To: rasmb at rasmb-email.bbri.org
Subject: [RASMB] Centrifugal force and protein dissociation


----------------------------------------------------------------------------
------
The older archived RASMB emails can be found at:
http://rasmb-email.bbri.org/rasmb_archives
and current archives at
http://rasmb-email.bbri.org/pipermail/rasmb/
Search All the Archives at:
http://rasmb-email.bbri.org/rasmb_search.html
----------------------------------------------------------------------------
------

Dear colleagues:

Currently, our lab has tried to determine the heterogeniety of an old
protein sample.  The protein, if folded correctly and behave well, should
remain tetrameric for a long time.  However, by sed. Vel. we found that
the protein sample is the mixture of monomer, dimer, and tetramer.  The
fit was done by DCDT (by John Philo).  Here are a few questions that I
would like to ask:

1)  Can centrifugal force tear multimeric protein apart?  This protein, in
particular, has a central channel where water can go through.

2)  If there is monomer-dimer-tetramer equilibrium, can we trust
sedimentation velocity data in terms of their relative amount?

3)  What is the reliability or error from the sed. Vel. fit?

We are very confused whether the fact that we observe
monomer-dimer-tetramer is true or is just an artifact.  I would greatly
appreciate your insight, thought, or suggestion in light of these
concerns.

Sincerely yours,
Songpon Deechongkit
PhD Candidate
Jefferey W. Kelly Lab
The Scripps Research Institute
La Jolla, California


_______________________________________________
RASMB mailing list
RASMB at rasmb-email.bbri.org
http://rasmb-email.bbri.org/mailman/listinfo/rasmb