<html>
<body>
Hi Mitra,<br><br>
Regardless of the type of analysis you do (whole boundary fitting or
weight average S points) an aggregate species will affect the
analysis. How it affects the analysis depends on the S value and
kinetics of the aggregate.<br><br>
We have had aggregates that are so large that they are essentially
pelleted by the time the rotor reaches final speed before the first scan
and no new aggregates form during the time span of the experiment.
These have very little effect on the analysis except you must use the
concentration from the centrifuge's absorbance or refractive increment
measurement instead of the believed loading concentration. At times
it seems that trying to remove such aggregates before loading the cells
(by pre-spinning and pelleting them) disturbs the dialysis equilibrium,
and so for interference optics experiments, it is best to just let them
pellet during the run rather than trying to get rid of them before
loading the analytical cells.<br><br>
In all other cases, the best way to deal with irreversible aggregates is
to fit only the data that does not include the aggregates. The
early scans near the meniscus always include the aggregates and should
not be used in the fit or the determination of weight average S.
For best separation, choose the data scans where the smaller species of
interest are close to the bottom of the cell and the aggregate is
pelleted. If you use Sedanal, you simply choose a late
range of scans. We jokingly call this 'purifying' our sample.
It is a valid strategy when the aggregate boundary does not overlap the
boundary of the samples of interest. DCDT+ actually lets you set an
S value range to fit, though you still have to avoid early scans where
diffusion dominates and the boundaries almost always overlap on the S
scale. In the recent AUC workshop, John Philo demonstrated on some
data that fitting the whole range resulted in an incorrect determination
of Molecular Weight while fitting the restricted range ignoring the
aggregates gave the right answer. Presumably, if you get the wrong
Molecular Weight you would also get the wrong weight average S.
<br><br>
In one case, we successfully modeled a monomer, dimer, tetramer, octomer
equilibrium in which one sample had aggregate going faster than
octomer. In this case it was necessary to hold the S value of the
octomer fixed (to a value from the sample when it was fresher and had no
aggregate) or the S value for the octomer went too high to try to best
fit the aggregate. There seemed to be too much overlap between the
octomer and the aggregate to just choose the right scans, but when we
held the octomer S value fixed, then it seemed to ignore the aggregate
and you saw the bump in the residuals at higher S.<br><br>
If the aggregates are forming during the experiment in a measurable
amount, you can never choose a set of scans that do not include them and
in that case have to add them to your model. In this case, I have
heard of people modeling the 'aggregate' equilibrium like Jack Correia
models Tubulin polymerization to tease out tubulin to sample binding
constants. I am not familiar with anyone doing weight average S
isotherms with a background of aggregate formation, though it does not
sound impossible it probably requires that some one of the theorists have
worked out the equations for this kind of fit (or maybe you would be
brave enough to write out the equations yourself).<br><br>
With any of the major AUC software packages (SEDANAL, DCDT, SEDFIT) you
can simulate a set of experiments. I know that with the model
editor in SEDANAL you can model any of the situations I described
(reversible or irreversible aggregation) with the model editor, you can
then investigate yourself what effect an aggregate like what you see has
on the analysis: just simulate data sets and then try out your
analysis strategy and see if it returns the values set in the
simulation.<br><br>
Hope this helps.<br><br>
Dr. David B Hayes<br>
Analytical Ultracentrifugation Research Laboratory<br>
Boston Biomedical Research Institute<br>
64 Grove St.<br>
Watertown, MA 02472<br>
617-658-7738<br><br>
<b>Boston Biomedical Research Institute... Today's Research for
Tomorrow's Health.<x-tab> </x-tab><br>
<i>Please visit us at
<a href="http://www.bbri.org/" eudora="autourl">www.bbri.org<br><br>
<br><br>
</a></i></b>At 04:04 PM 5/29/2007, you wrote:<br>
<blockquote type=cite class=cite cite="">Hi Folks<br>
I have a question (s) - how does the aggregation (and maybe the
subsequent loss)<br>
of a protein sample during SV affect the Kd of the monomer-dimer
equilibrium?<br>
Either by direct fitting of the boundary or considering the average<br>
sedimentation coefficient. Is there any way around this?<br><br>
Mitra<br>
-- <br>
Mitra S. Rana<br>
Graduate Student<br>
Institute for Cellular and Molecular Biology<br>
2500 Speedway, UT-Austin<br>
Austin, TX-78712<br>
_______________________________________________<br>
RASMB mailing list<br>
RASMB@rasmb.bbri.org<br>
<a href="http://rasmb.bbri.org/mailman/listinfo/rasmb" eudora="autourl">
http://rasmb.bbri.org/mailman/listinfo/rasmb</a></blockquote></body>
</html>