From: STEVE HARDING <sczsteve@szn1.agric.nottingham.ac.uk>
  To  : rasmb@bbri.harvard.edu
  Date: Mon, 28 Apr 1997 16:56:22 GMT

Experiments in Glycerol: Summary of views

Hello All

A fortnight ago I asked if anybody had experience of experiments in 
glycerol.  The response was fantastic and just confirms what a 
useful sytem this RASMB is.  Many thanks to all (Marc, Don, Jim, Tom, 
Sue and Arthur).  For the file, and if anyone else has course to use glycerol
 systems in the future, here is a summary of the correspondence, starting with
 my initial request

Cheers

Steve


>From Steve Harding:
"We've been asked to characterise a multi-subunit DNA binding protein -
the RPO-protein which is extremely unstable even at 4 deg. The guys
won't let it out of glycerol (20% minimum).  Is there anybody out
there with experience at working under these viscous conditions which
seem to me extremely unfriendly to the u/c operator?  Higher
temperature to reduce the viscosity is of course disallowed! Has there
been anything published on BSA or something?  I guess
 velocity experiments are ruled out and equilibrium will take ages. 
 Arthur (Rowe) has pointed out there may be vbar & dialysis problems 
too".  

The responses (Marc Lewis, Don Winzor, Jim Cole, Tom Laue, Sue 
Bethel, Arthur Rowe):


Marc Lewis:
"We have been working with human DNA polymerase-beta (beta-pol) and AP
Endonuclease (APE) in 30% glycerol.  The time to equilibrium at 4
degrees C. was 10 days for a 4 mm column at 15000 RPM.  As far as vbar
was concerned, the calculated value of vbar for APE using Perkins'
consensus values and the calculated value of rho for the buffer gave a
calculated value of (1 - vbar*rho) in quite good agreement with the
value measured in the XL-A using the value of M calculated from the
amino acid composition. We could not make this comparison for beta-pol
since it is in a monomer-dimer equilibrium.  As a general rule,the obvious 
suggestions of using the shortest columns that give adequate resolution and having a lot of patience is
about all you can do.  I have also found that the providers of protein
samples "needing" glycerol frequently tend to grossly overestimate the
concentration of glycerol needed to provide stability. "

Don Winzor:
"I can't do much about the time it takes to reach equilibrium; but I
can handle the analysis.  Yes, vbar goes up in smoke to become the
wretched (phi')app thing that Casassa and Eisenberg dreamed up and
Timasheff religiously measured.  However, the trick is to avoid the
measurement of that parameter altogether.  Use the buoyant molecular
weights of protein and DNA in 20% glycerol to define psiS(r) and
psiA(r) and use Helmut's (Helmut Colfen)  program to obtain a value of K.  This value
will be an apparent value that incorporates the nonideality arising
from excluded volume interactions with 20% glycerol.  However, by
repeating the experiment in a series of glycerol concentrations, one
could try extrapolating the apparent K value to zero glycerol
concentration.  Do they really need 20%; or can they come down to 15%
or even 10%?"


Jim Cole:
Last year we also had to do some centrifugation work in up to 20%
glycerol with cytomegalovirus protease.  At 20 C, the viscosity was
not too much of a problem for equilibrium and velocity work. For a
variable temperature equilibrium study we worked at 10% glycerol, and
for a 3 mm column height equilibrium required up to two days at 5 C.
As you  suggested, glycerol does affect v-bar.  Timasheff published a
study of glycerol effects a while ago (Gekko and Timasheff
Biochemistry 1981, 20, 4667-4676). The dependence of v-bar on
[glycerol] was found to be similar for four different proteins. From
their data we calculated an average correction factor for v-bar: Delta
v-bar/Delta % volume glycerol = +3.33 X 10-4. Hope this helps"

Tom Laue:
"We worked with an interacting system in 40% glycerol. The solution is
like honey, so there is no hope for velocity work. Your only hope is
to use short solution columns and wait for equilibrium... forever. We
found with 0.75 mm columns that 24 hrs or more were required (and we
were at 20 C). Be sure to use a program like Match (at the RASMB site)
to determine equilibrium- you can be fooled easily by subtracting
scans.

Arthur is right about vbar, and I don't know of anyone who has done
systematic work on that (although Timasheff may have). Let me know if
you find anyone who has any idea about the general trend for vbar in
glycerol.

Often the high dipole moment of DNA-binding proteins causes
aggregation. This is usually reported as 'instability' by the persons
working with the protein. I have found that high salt sometimes makes
these balky systems behave better"


Sue Bethel:
"I have been working on a protein in 30% glycerol and although slow,
have not have any real problems with it.  The data from the AUC
correlated extremely well with that obtained from gel filtration.  I
used the vbar calculated and in our case it did not seem to have
problems. For equilibrium runs at 4C I gave it a 1 hour blast at 40K
before slowing it down to the speed I wanted.  This trick (courtesy of
Tom Laue) took about 12 hours off the equilibration time (It still
took about 20 hours to reach eq).  We did not see any evidence of the
glycerol separating out".

Arthur Rowe:
"Obviously there will be a major effect on time to attain equilibrium: 
the approach to equilibrium methods might be advantageous if you want
to avoid 1 week runs !  Effect on vbar could be major, inasmuch as
bound solvent is likely to be very untypical of bulk solvent, and a
DDM expt is probably hard to do (how good would dialysis be ?), though
possible".