[RASMB] Gilbert-Jenkins and Multisignal-SV for stepwise association of complex

[Ian Berke]

Dear all,

I am a relative newbie to AUC, but am interested in generating a global
analysis of SV titration data to get an estimate of the Kd and cooperativity
of my system.  I have an RNA-binding protein that runs at 4.4S and binds to
a 20bp dsRNA of 2.6S.  I've done a titration of 1uM dsRNA with increasing
amounts of protein, spun at 45K rpm and monitored at 260 and 280nm (and IF,
but was very noisy).  I estimated that the Kd was ~< 1uM based on EMSA
results.  I kept the RNA concentration constant near the estimated Kd since
the A260 would get too high otherwise. Attached is an overlay of the
resulting c(s) (from 260nm data).  You can see the 2.6S RNA peak decrease as
protein is added with a 5-6S (concentration-dependent) species appearing as
well as a 7.8S species (whose position remains pretty constant over the
concentration range tested).  The concentration dependent S-value first
increases from 5.7 to ~6S, then decreases as more protein is added.  I
interpret this as a "fast" kinetic reaction (koff>10^-4) of a 1:1 complex
and a "slow" dissociating 2:1 complex (protein:RNA).  Is that a reasonable
interpretation?  I'd thought to use Gilbert-Jenkins Sfast/Sundisturbed and
Pop-fast/Pop-undist. isotherms to analyze the data, but couldn't find
anything in the literature for a 2:1 complex.  Is GJ approach valid/feasible
for anything more than 1:1 system?  I'd like to generate isotherms from this
data to fit with sedphat in a A+B-->AB+B-->ABB model.

Also, I tried doing a multi-signal SV analysis to get the stoichiometry of
the 7.8S complex (280 and 260nm), and it roughly works (1.75:1, have to
better refine extinction coefficients-protein alone was only estimated), but
the c280(s) and c260(s) peaks were separated by ~0.7S for the region of the
complex (7.4S and 8.1S see other attached figure).  I was thinking this
might have to do with the rate of data aquisition for each scan (older
machine with 3 cells being scanned at 280nm and 260nm, so 260nm scan is some
minutes after 280nm).  In all, there were only 15-16 scans in which the fast
boundary was observed so maybe that has an effect too.

Finally, if I do a multi-signal SV analysis on each concentration point to
get an estimate of each species in the concentration-dependent reaction
boundary, the free/undisturbed species and the final complex, could I
generate isotherms with that to analyze with the A+B-->AB+B-->ABB model?

Any advice or help would be appreciated, even if it's just to be told I'm a
raving lunatic for attempting this.

Thanks,
Ian
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.rasmb.org/pipermail/rasmb-rasmb.org/attachments/20100722/9f7a4725/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: titration-cs.pdf
Type: application/pdf
Size: 54120 bytes
Desc: not available
URL: <http://list.rasmb.org/pipermail/rasmb-rasmb.org/attachments/20100722/9f7a4725/attachment.pdf>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: cell3-mssv-float.png
Type: image/png
Size: 184745 bytes
Desc: not available
URL: <http://list.rasmb.org/pipermail/rasmb-rasmb.org/attachments/20100722/9f7a4725/attachment.png>