[RASMB] viscosity issue in sedimentation velocity measurements of complex formation

[Arthur Rowe]

Greeting Donald, and everyone

You sound to have a nice system there. Having a small protein binding  
to a (much) larger one is easier to analyse than where the two proteins  
are comparable in size, although at the low concentrations you are  
working at, and on the presumption that your proteins are more or less  
'globular', c-dependence of the s values of the species ought not to be  
a serious problem.

To see whether you need to bother about the 'viscosity of slow  
component effect' let's have a look at a 'worst case' scenario of the  
faster component having to sediment through concentrations of up to 1  
mg.ml of the slow component. With wide separation of s values, you can  
take the contribution of the viscosity of the slow component as being  
simply additive to that of the solvent (buffer). If the intrinsic  
viscosity of the slow component is around 5 ml.g, which is a fair  
guess, then at the most the total viscosity of (solvent + slow  
component) will be around 0.5% greater than that of (solvent alone), at  
the maximum concentration of 1 mg.ml for the slow component. Note that  
in the actual case, it will be a bit lower than this, because of the  
concentration being less than 1 mg.ml because of  some of it being  
bound to the fast component. And a tiny bit due to radial dilution.

So, on the face of it, you might think that the whole effect can be  
ignored. But its not as simple as that. The % change in the s value of  
the fast component which you are looking for, and have indeed detected,  
is only around 10%. So the real error in ignoring the viscosity  
contribution of the slow component is order of 0.5% in 10%, i.e. you  
will be under--estimating the change in the s value of the fast species  
by around 1 part in 20. This issue could be pursued both theoretically  
and experimentally at some length. But what I would say is that if you  
simple increase the s value of the fast species by an amount  
proportional to the amount of slow component contribution, the  
correction, although not perfect, will be much better than making no  
correction at all.

That being said, however, maybe you should think of moving on from  
using SVEDBERG as your default software. It's not that a single species  
fit in SVEDBERG will give you a faulty vale, as compared to using  
SEDFIT, for example, but there is much more you can do with the latter.  
Although from an experimental point of view, if your aim is to  
quantitate the amount of slow component which gets bound, you could do  
this better, even sticking to SVEDBERG, by going into the far u/v to  
pick up that slow component (I bet there is a peak around 230 nm or so)  
and measuring the total area found at various c values, As compared to  
control (no fast species present). You would of course spin at a speed  
high enough to pellet the fast species, and only take for analysis  
scans after that was achieved. But I'd still want to use SEDFIT myself.

Regards to you and all

Arthur

************************************************************************ 
*******
Arthur J Rowe
Professor of Biomolecular Technology / Director NCMH Business Centre
School of Biosciences
University of Nottingham
Sutton Bonington
Leics LE12 5RD

TEL:  0115 9516156
FAX:  0115 0516157
************************************************************************ 
*******

On Jan 21, 2010, at 18:51, Donald W. Pettigrew wrote:

> Hello, Rasmb
>  
> I have a question about viscosity corrections for s20,w due to varying  
> concentrations of protein.
>  
> We are using sedimentation velocity to determine the binding constant  
> for association of an 18.1 kDa protein with a 224kDa protein.  The  
> concentration of the larger protein is maintained constant at 0.3  
> mg/mL (about 5 uM subunits for the tetrameric protein) and the  
> concentration of the smaller protein is varied from 0-1 mg/mL (about  
> 50 uM).  The binding stoichiometry is known to be 1 small protein per  
> subunit of the larger protein (4 per tetramer).  The dissociation  
> constant is known from independent studies to be about 5 uM.
>  
> We run sedimentation velocity experiments at 28,500 rpm, using  
> absorbance optics at 280 nm, standard aluminum double sector cells.   
> We have the great fortune that the smaller protein is almost entirely  
> transparent at 280 nm.
>  
> At this juncture, we analyze the data by using SVEDBERG to obtain a  
> single value for s20,w.  We see the s20,w increase from about 11S to  
> about 12+S as the concentration of the smaller protein is increased  
> 0-1 mg/mL.  This increase is about what we expect.  We used the same  
> corrections for s20,w at all concentrations of the smaller protein,  
> although the solution viscosity should increase with increasing  
> concentrations of the smaller protein.  Thus, we are underestimating  
> the s20,w at the higher concentrations of the smaller protein.  
>  However, I have no feel for the extent of the error.
>  
> We are sort of following the paper: Dam et al (2005) Biophys J  
> 89:619-634 from Peter Schuck, although we have not yet gotten to the  
> elegant analysis part.  I am not sufficiently familiar with the  
> computational methods that are described in the paper to see whether a  
> correction for viscosity is there.
>  
> We can determine empirically the viscosity effect on sedimentation of  
> the larger protein by using mutants of it that we know do not bind the  
> smaller protein.  However, I wonder whether this is sufficient or are  
> there better ways to correct for the viscosity differences?  If so,  
> what are they?  Does anyone on the listserv know about this issue?
>  
> Thanks for your consideration of this lengthy email.
>  
> Dr. Donald W. Pettigrew
> Department of Biochemistry & Biophysics
> Texas A&M University
> College Station, TX 77843-2128
>  
> phone: 979-845-9621
> fax: 979-845-9274
>  
> "Nearly all  men can stand adversity, but if you want to test a man's  
> character, give him power"
> A. Lincoln
>  
> _______________________________________________
> RASMB mailing list
> RASMB at rasmb.bbri.org
> http://rasmb.bbri.org/cgi-bin/mailman/listinfo/rasmb
>

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/enriched
Size: 8108 bytes
Desc: not available
URL: <http://list.rasmb.org/pipermail/rasmb-rasmb.org/attachments/20100122/be54b283/attachment.bin>