[RASMB] combining DLS and sedimentation equilibrium to get molecular shape

[Peter Schuck]

Hi Dave,

In theory I think this should work, but in practice I believe the validity 
of this will depend on how mono-disperse your sample is.  It is quite 
possible, if contaminations by small Mw species are not resolved in 
sedimentation equilibrium, that the molar mass estimate could be too 
low.  On the other hand, small impurities would go undetected by DLS, but 
DLS would be highly sensitive to larger Mw impurities (which may not be 
detected by sedimentation equilibrium if they sediment very close to the 
bottom of the solution column).  Taken together, I think the danger is that 
the frictional ratio values would tend to be overestimated.

One can use SEDPHAT to globally model SE and DLS and see if both jointly 
are consistent with a single species model. (In fact, you can directly 
model s-values and molecular weights for one or more species globally to 
the autocorrelation functions and the SE profiles.)

I appreciate you have already done the experiments, and obviously there's 
not enough material to redo.  But on the experimental side, it sometimes is 
a good idea to shake up a SE column and rerun it as a short-column 
SV.  That would give at least some additional information.  Alternatively, 
the following has worked very well for me:  After achieving SE, I carefully 
stopped the rotor and took the cells out preventing mixing of the solution 
column as much as possible.  (When working with proteins that absorb in the 
VIS, it is apparent that you can take the cells out of the rotor without 
completely remixing the solution).  Then, I was drawing the DLS sample with 
a syringe from the middle of the SE solution column.  That, at least, give 
a sample for DLS that is free of larger contaminants (and it beats 
filtering or spinning it down quickly to get rid of dust).

Best,
Peter



At 10:22 AM 3/15/2006, David Lee wrote:
>Dear colleagues,
>I am writing to enquire about the validity and pitfalls of modeling
>hydrodynamic shape by combining sedimentation equilibrium data with
>dynamic light scattering data. I didn't have enough protein for
>sedimentation velocity experiments so after performing sedimentation
>equilibrium experiments, whcih showed the complex to behave as a
>single ideal species, I measured the hydrodynamic radius by dynamic
>light scattering, then calculated the frictional coefficient f. I
>combined this number with the molecular weight obtained from the
>equilibrium experiment to calculate s* which I entered along with
>the other data into SEDNTERP to model the dimensions of the protein
>as a prolate ellipse. Is this a valid way of doing hydrodynamic
>modeling? I'm worried that I'm overlooking something. Thanks in
>advance for your input on this matter.
>
>Sincerely,
>Dave
>
>_______________________________________________
>RASMB mailing list
>RASMB at rasmb.bbri.org
>http://rasmb.bbri.org/mailman/listinfo/rasmb