[RASMB] Analysis of sed. equilibrium data

[Patrick Brown]

Dear fellow AU users:
I have attempted to determine if my protein self-associates by sedimentation equilibrium.
 
The experiment was performed at three protein loading concentrations (0.1, 0.6, 1.2 mg/mL) and spun at 2 rotor speeds (28k, 36krpm).  The system was determined to be at equilibrium by overlaying the abs. vs. radius profiles collected at 22 hours and then 1 hour later.  The data was analysed using WinNonLin (very user-friendly) first as a single species then as various association models- which did not improve the fit (based on SRV).
 
For the single-species analysis, deltaY1, LnA1 and sigma were allowed to float.  The square-root-variance (SRV) was 0.005 or better, however the residuals appeared slightly skewed (drifting upward then back downward) but not significantly.  I have made two observations based on that analysis of the data that I am hoping someone might be able to explain.
 
1) First, the molecular weight calculated from the range of sigma-values returned during the fit of each data set is lower than what is calculated from the amino acid composition by 9.4%.  I thought I read somewhere that a 5% difference in the MW was acceptable- 
[I determined the vbar and rho from sednterp]
 
2) second, the sigma-value obtained from the fit is concentration dependant.  The sigma value decreases with higher loading concentration.
 
Do these observations sound like I am making measurements under non-ideal conditions?
[My protein is 20kDa, a pI of 8.6 (charge 1.6 at pH 7.5), and is in 10 mM Tris-HCl, 100 mM NaCl]
 
Any comments are greatly appreciated.
Rick

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.rasmb.org/pipermail/rasmb-rasmb.org/attachments/20041105/713cc555/attachment.htm>