[RASMB] vbar

Wed Jun 7 10:48:51 PDT 2006

Ewa,

Yes I think you have missed a key point. The point was NOT to estimate
stoichiometry based on the actual Rh values and some calibration curve of Rh
versus M. Rather, Peter's proposal was to get the molar mass by taking the
ratio of the sedimentation coefficient (from SV of course) to the diffusion
coefficient (measured directly by DLS), using the Svedberg equation. 

For that approach in order to distinguish 21-mer from 24-mer you need only
measure the diffusion coefficient to ~8% accuracy. (The sedimentation
coefficient will easily be accurate to 0.5% or better.) The drawback to this
approach is that it still requires an accurate vbar value to get an accurate
true molar mass.

But actually yes I do find that the precision of the diffusion coefficients
from DLS (and therefore Rh values too) can be about 2%. And as Hans
Schoenfeld said, for homogeneous samples the diffusion coefficients from DLS
do generally agree quite well with those from sedimentation velocity.

But overall I would say static light scattering (after SEC or using Allen
Minton's approach) is probably the easiest and most reliable approach to
resolving this discrepancy.

John

-----Original Message-----
From: rasmb-bounces at rasmb.bbri.org [mailto:rasmb-bounces at rasmb.bbri.org] On
Behalf Of Ewa Folta-Stogniew
Sent: Wednesday, June 07, 2006 9:07 AM
To: Schoenfeld, Hans-J.; Allen Minton; rasmb at server1.bbri.org
Subject: RE: [RASMB] vbar

Hi all,

my point was that for such big oligomers (i.e.
composed of many units) one should go after molar mass
rather than size to distinguish 21-mer vs. 24-mer. 
The 24-mer differs by ~14.3% from 21-mer in mass while
it differs only by ~5.8% in size.  Thus, unless one
can measure Rh with accuracy better than 2% (it this
doable?), it seems virtually impossible to assign the stoichiometry with
confidence.... 

Am I missing something?

Ewa 

--- "Schoenfeld, Hans-J."
<hans-j.schoenfeld at roche.com> wrote:

> Static light scattering may be simpler as it saves
> the sedimentation experiment.
> However, to feel confident I would do both. Both
> methods depend exclusively on absolute measurements
> of parameters (M vers. D and s), should therefore
> result in similar numbers for M and would in the
> best case confirm each other...
> 
> Hans-Joachim Schönfeld.
> 
> -----Original Message-----
> From: rasmb-bounces at rasmb.bbri.org 
> [mailto:rasmb-bounces at rasmb.bbri.org] On Behalf Of Allen Minton
> Sent: Wednesday, June 07, 2006 3:21 PM
> To: rasmb at server1.bbri.org
> Subject: RE: [RASMB] vbar
> 
> 
> I second Ewa's suggestion that static, not
> dynamic, light scattering seems to be the method 
> of choice here, provided that you have an 
> accurate determination of concentration and 
> refractive increment.  The ratio of scattering 
> intensity to concentration is proportional to 
> molar mass and can easily be determined to within 
> a few percent in less than 15 min.  We do this 
> routinely.  The only caveat is that special 
> attention must be paid to sample preparation: for 
> best results the sample must be filtered through 
> a 0.02 micron filter and degassed by centrifugation
> just prior to measurement.
> 
> Allen Minton
> 
> At 03:32 AM 6/7/2006, you wrote:
> >Ewa and Peter,
> >I agree with both of you, that DLS is not
> >precise in the characterization of single
> >components of polydisperse samples. However, I 
> >also agree with Peter that the method is far 
> >better doing in the analysis of monodisperse 
> >samples. To my experience, diffusion 
> >coefficients as obtained by DLS are precise and 
> >highly reproducible (determination is absolute 
> >and is mainly based on measurement of time which 
> >can be done with very high precision...).
> >Our experience was that the strategy to combine 
> >s from AUC with D from DLS resulted in molecular 
> >masses that were very close to values as 
> >obtained from sedimentation equilibrium runs and 
> >I therefore recommend to give it a trial (see 
> >table in: Schoenfeld, Poeschl, Mueller: 
> >"Quasi-elastic light scattering and analytical 
> >ultracentrifugation are indispensable tools...", 
> >Biochemical Society Transactions, 26, pp.753-758
> (1998)).
> >Best regards,
> >Hans-Joachim.
> > >============================================
> > >Dr. Hans-Joachim Schönfeld
> > >F. Hoffmann-La Roche Inc.
> > >PRBD-E, B93/5.44
> > >CH-4070 Basel
> > >Switzerland
> > >
> > >Tel. (+41) 61 688 28 95
> > >Fax. (+41) 61 688 90 60
> >mailto:hans-j.schoenfeld at roche.com
> >
> >
> >-----Original Message-----
> >From: rasmb-bounces at rasmb.bbri.org 
> >[mailto:rasmb-bounces at rasmb.bbri.org] On Behalf Of
> Peter Schuck
> >Sent: Tuesday, June 06, 2006 11:08 PM
> >To: Ewa Folta-Stogniew; Chad Brautigam
> >Cc: rasmb at server1.bbri.org
> >Subject: Re: [RASMB] vbar
> >
> >
> >Ewa,
> >
> >I agree with you that one maybe should not use a
> size-distribution
> >model in that case.  However, this limitation is
> not one of DLS, but
> >the analysis method used.
> >
> >  From the sedimentation velocity, it should be
> possible to assess very
> >well how monodisperse the peak is.  If it appears
> to be pretty much a
> >single species (without very large aggregates, or
> those could perhaps
> >be filtered out), then one can do the DLS analysis
> nicely with a single
> >species model.  For example, both SEDFIT and
> SEDPHAT allow you to use
> >those discrete models and fit DLS autocorrelation
> functions, and I
> >believe some instrument makers also supply software
> that can give a
> >single diffusion coefficient, rather than a
> distribution.  That's what
> >I would try to combine with s.
> >
> >Peter
> >
> >
> >At 04:32 PM 6/6/2006, Ewa Folta-Stogniew wrote:
> > >Peter and Chad,
> > >
> > >I maybe too pessimistic, but I do not believe
> that DLS
> > >has enough precision to distinguish 21 vs. 24 configuration. The 
> > >standard fitting protocol for
> Rh determination is
> > >for Gaussian distribution of sizes with 15% SD,
> which would encompass
> > >sizes for both configurations making them
> virtually identical.
> > >
> > >Static LS for MW determination maybe an option
> given
> > >the experiment is done with accuracy at least +/-
> 3%.
> > >
> > >
> > >Ewa
> > >
> > >--- Peter Schuck <pschuck at helix.nih.gov> wrote:
> > >
> > > > Hi Chad,
> > > > if the oligomer is relatively pure, would it
> be
> > > > possible to do
> > > > DLS? Perhaps taken together with s, that would
> give another quick
> > > > way of confirming M, instead of equilibrium
> sedimentation.
> > > > I don't think enclosed
> > > > solvent adds to the buoyant molar mass,
> therefore it
> > > > would not contribute
> > > > an error to the vbar. However, due to error
> > > > propagation from vbar to the
> > > > (1-vbar*rho) term, small errors in vbar are
> > > > amplified; my feeling is that
> > > > this usually could account for a few percent uncertainty (i.e. 
> > > > in the ballpark of 1/24).
> > > > Peter
> > > >
> > > > At 02:45 PM 6/6/2006, you wrote:
> > > > >Hello, All,
> > > > >
> > > > >Sorry if this is a rudimentary question. I
> have
> > > > been using velocity
> > > > >sedimentation to examine the oligomeric
> states of a
> > > > protein and
> > > > >mutants thereof. Some mutants are trimers,
> and the
> > > > molecular weight
> > > > >estimates given by sedfit (either a c(M)
> > > > distribution or a discrete
> > > > >species model) are very reasonable. However,
> based
> > > > on a crystal
> > > > >structure, we expect the wild-type to be a
> 24-mer.
> > > > Sedfit
> > > > >consistently underestimates the molecular
> weight (
> > > > I get something
> > > > >more akin to 21-mer).
> > > > >
> > > > >I assume that there are at least to
> possibilites
> > > > here:
> > > > >
> > > > >1. The crystal structure is wrong, and the
> thing
> > > > really is a 21-mer
> > > > >in solution.
> > > > >
> > > > >2. The vbar calculated by Sednterp is
> inaccurate-
> > > > it is not
> > > > >accounting for the fact that some of the
> volume of
> > > > the 24-mer is not
> > > > >taken up by protein, but by solvent. The vbar
> is
> > > > therefore
> > > > >significantly too low, with adverse effects
> on the
> > > > MW calculation.
> > > > >
> > > > >Does anyone know if there is a more accurate
> way to
> > > > estimate vbar in
> > > > >cases of large macromolecular assemblies? Can
> our
> > > > crystal structure
> > > > >help us out in any way?
> > > > >
> > > > >BTW, yes, I know that sed. equilibrium might
> be the
> > > > preferred
> > > > >approach in this case, but instrument time is
> > > > limited at the moment.
> > > > >
> > > > >Thanks,
> > > > >Chad
> > > > >
> > > > >
> > > > >==================================
> > > > >Chad A. Brautigam, Ph.D.
> > > > >Research Scientist
> > > > >The University of Texas
> > > > >Southwestern Medical Center at Dallas
> > > > >5323 Harry Hines Blvd.
> > > > >Dallas, TX 75390
> > > > >Office:  (214) 645-6384
> > > > >Fax:      (214) 645-5383
> > > > >
> > > > >
> > > > >
> > > > >
> > > >
> >_______________________________________________
> > > > >RASMB mailing list
> > > > >RASMB at rasmb.bbri.org
> http://rasmb.bbri.org/mailman/listinfo/rasmb
> > > >
> _______________________________________________
> > > > RASMB mailing list
> > > > RASMB at rasmb.bbri.org
> http://rasmb.bbri.org/mailman/listinfo/rasmb
> > > >
> > >
> > >
> >
> >__________________________________________________
> > >Do You Yahoo!?
> > >Tired of spam? Yahoo! Mail has the best spam
> protection around
> > >http://mail.yahoo.com
> >_______________________________________________
> >RASMB mailing list
> >RASMB at rasmb.bbri.org
> http://rasmb.bbri.org/mailman/listinfo/rasmb
> >_______________________________________________
> >RASMB mailing list
> >RASMB at rasmb.bbri.org
> http://rasmb.bbri.org/mailman/listinfo/rasmb
> _______________________________________________
> RASMB mailing list
> RASMB at rasmb.bbri.org http://rasmb.bbri.org/mailman/listinfo/rasmb
> _______________________________________________
> RASMB mailing list
> RASMB at rasmb.bbri.org
> http://rasmb.bbri.org/mailman/listinfo/rasmb
> 

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 
_______________________________________________
RASMB mailing list
RASMB at rasmb.bbri.org http://rasmb.bbri.org/mailman/listinfo/rasmb