From: Gunther.Kern@astrazeneca.com
  To  : rasmb@bbri.org
  Date: Sat, 23 Oct 1999 02:13:29 +0200

sedimentation runs at high salt

Dear mail group members. 
I would like to thank all who took the time to respond quickly (Walter
Stafford, Christine Ebel, Karen Flemming, Tom Laue, Emory Braswell, John
Philo, Marc Lewis, Arthur Rowe). All of the comments were very helpful.
As a short overview I would like to list  some of the comments that were not
communicated through the rasmb mail address and where I learnt something
from.

Tom Laue wrote:

The effects on diffusion will be exactly cancelled by the effects on
sedimentation. The most probable cause of trouble for you is a shift in
vbar at high salt. In general, vbar goes down. There is some work on this
in the literature-
P. Tuengler, G. L. Long, and H. Durchschlag. Calculated Molecular Weight of
Proteins in High Ionic Strengths: Contribution of the Apparent Isopotential
Specific Volume. Anal.Biochem. 98:481-484, 1979.

Since I do not see variation in MW when shifting concentrations or speed
Toms comment was dead on the problem and I was amazed how much vbars can be
affected.

Christine Ebel took the time to do some calculations and also sent me useful
references.
Here is a short summary of her part of the discussion:

I have performed equilibrium sedimentation runs at high salt. All your
results can be explained taking into account the hydration of the protein,
due to the fact that solvated water is visible when the density of the
solvent is different from the density of water.

If I do not make mistakes:
>From the informations you give, assuming v2bar for the protein=0.73, you
measured an apparent molar mass of about 41000(1-1.158X0.73)=6341.
Mapp=M ddc
M=molar mass of the protein part.
ddc= density increment (due to the protein and pertubated solvent)
If the protein is 60KDa, I calculate for the density increment ddc a value
of 6341/60000= 0.106

ddc= (1-rho.V2)+KSi1 (1-rho. V1)
V2= partial specific volume of the protein= 0.73
V1=partial specific volume of water =1
rho= solvent density= 0.73
Ksi1= water binding parameter
(1-rho.V2)=0.155
(1-rho. V1)=-0.158
ksi1=(ddc-(1-rho.V2))/(1-rho. V1)= (0.106-0.155)/(-0.158)=0.049/0.158=0.31g
of water by g of protein.

In general, this parameter IS NOT the water bound by the protein. It
contains also the interactions with salt (in the case of the very acidic
protein halophilic malate dehydrogenase that I have studied a lot, these
interactions with salt are shown to be not negligeable). In a structural
model in which the protein binds B1 g of water and B3 g of salt:
KSI1=B1 - B3/w3
W3 is the composition of the solvent expressed in gram of salt by gram of
gram of water.

Your results are thus compatible with a 60000 protein solvated with 0.31g/g
water and no salt, which is a very usual value (but other processes of
interactions are not excluded: which more water and some salt associated).

All the theoritical work has been done by H. Eisenberg. I explain some of
the concepts and give the references in my papers in Colloid & polymer
Science.
Ebel C., Characterization of the solution structure of halophilic proteins.
Analytical ultracentrifugation among complementary techniques (light,
neutron and X-Ray scattering, density mesurements),Progress.  Colloid &
polymer Science, 99 (1995) 17-23.
Kernel B., Zaccai G., Ebel C., Determination of molal volumes, and salt and
water binding, of highly charged biological macromolecules (tRNA,
halophilic protein) in multimolar salt solutions, Progress. Colloid &
polymer Science 113 (1999) 168-175.

John Philo wrote:

If you are talking about the fact that addition of KCl will alter the
viscosity (which will alter the time to equilibrium but not the
equilibrium distribution), according to SEDNTERP the viscosity of 3.2
M KCl is less than 2% different from that of water.

Certainly the high salt might affect the vbar, but I strongly doubt
such a change would be large enough to explain a 30+% drop in apparent
mass. It seems to me the most likely explanation for your result is
that there is significant dissociation into monomers.

Karen Flemming wrote:
D should have no effect on your molecular weight IF you are at
equilibrium.

It's more likely that you are either
a) NOT at equilibrium
b) faced with a situation where you have preferential binding or exclusion
of
water or KCl.