If I can jump in a bit late on this one (my e-mail has been down for a few
days), I'd like to add a couple of points to the excellent earlier
responses from Greg Cauchon and Scott Leigh, particularly regarding the
single versus multi-angle issue and whether QELS (also known as dynamic
scattering, DLS, or photon correlation spectroscopy, PCS) in a flow mode is
really very useful.
Unfortunately I think our lab may have created some confusion on the
single/multi-angle issue. Because the review article that Jie Wen, Tsutomu
Arakawa and I wrote in Analytical Biochemistry on protein applications of
on-line classical scattering primarily discussed using only the data from
the 90 degree detector, I fear that this has sometimes been misinterpreted
as implying that a single angle is always sufficient. As Greg correctly
explained, it depends a lot on the size range you wish to cover and exactly
what you are trying to learn.
Many of the applications involve (1) simply trying to identify whether the
native state is a monomer, dimer, etc. or (2) trying to identify the mass of
any aggregates that may be present in order to identify whether those
represent dimers, trimers, or whatever (what we call "oligomer hunting"). In
this type of application you are basically counting monomer units and you
really only need enough accuracy to be sure you do not miscount. If you are
working with 15-30 kDa monomers such are typical for many recombinant
protein pharmaceuticals, then even fairly large oligomers are well below a
physical size where there is any significant angular dependence, and the
extra angles are not particularly helpful. On the other hand, if you are
working with monoclonal antibodies, their much higher mass and extended
conformation mean that even a dimer is getting up in the range where with
only a 90 degree detector you start to lose accuracy by neglecting the
angular dependence.
I am not convinced that dynamic scattering is very useful in a flow mode as
an HPLC detector. There is a widespread misunderstanding that dynamic
scattering also measures mass in an absolute sense. That is just plain
wrong! Dynamic scattering really measures diffusion coefficients. Those
diffusion coefficients can be used to calculate the Stokes radius
("hydrodynamic size"), which is the radius of a spherical particle that
would have the same diffusion coefficient. If, and only if, you are willing
to assume that your protein is spherical, those diffusion coefficients or
Stokes radii can be translated to mass, but this is obviously often a
terrible assumption. To put it another way, the assumption of spherical
shape to get mass from dynamic scattering is exactly the same assumption
that is made in using size exclusion chromatography to estimate masses based
on a column calibration with globular protein standards, a notoriously
inaccurate practice.
If you are going to use an SEC column in front of a flow DLS detector
anyway, you will already be obtaining information about the hydrodynamic
size from the elution position. When used with SEC the only thing an on-line
DLS detector really does, in my view, is save you the trouble of calibrating
the column in order to get a hydrodynamic size (and it's pretty darn
expensive for that purpose).
On the other hand, if you have a classical scattering detector used with SEC
then you get the mass independently (and absolutely) and you can use the
elution position to get information about conformation. For example we saw
two different elution positions for dimers of an antibody, representing two
different conformations of dimer (we think side-by-side versus end-to-end).
Notably without the light scattering data the early eluting dimer fraction
had been previously assumed to be a tetramer. This example is shown on the
Alliance Protein Labs web site at http://www.ap-lab.com/light_scattering.htm
Having just trashed DLS in flow mode, I must point out that DLS can
certainly be useful in a batch mode for detecting the presence of trace
amounts of aggregates. A strength of batch DLS is that it can cover an
enormous range of sizes, and because the signal strength grows
proportionally to mass large aggregates can be seen even at levels of 0.01%
or less. In our experience, however, while it is good at detecting
aggregates, it is rather poor at quantitating whether that aggregate
represents 1%, 0.1%, or 0.01% of the total sample, even using very expensive
equipment. Indeed, often when there is any significant amount of aggregation
the aggregate signal swamps that of the native form, and the native form may
not even be detected (i.e. this technique can miss 99+% of the sample
entirely!) Nonetheless, for comparing *relative* amounts of aggregation
(this sample is worse than that, or changes with formulation conditions,
aging, or stress) it can be quite handy. Precision Detectors does offer a
batch-mode DLS add-on to their flow classical LS detector which looks to be
a relatively inexpensive way to get both types of light scattering, but I
haven't used their instruments so I can't comment on the quality of the
data.
Lastly, with regard to the different vendors, I certainly would echo Scott
Leigh's comments regarding good support and honest answers from the Wyatt
folks. We've also seen just excellent reliability with the
MiniDawn---several years of daily use with zero down time and essentially no
problems with fouling of the cell. Greg Cauchon has noted that the cell
volume for the Wyatt detectors is pretty large, which is somewhat a
drawback, but we generally prefer to use these detectors with a Pharmacia
Superose or Superdex FPLC column (to minimize stickiness and loss of
aggregates to the column) so this hasn't been much of an issue for us.
'Hope this helps,
John Philo, Alliance Protein Laboratories
www.ap-lab.com
-----Original Message-----
From: Association of Biomolecular Resource Facilities
[mailto:abrf-request@aecom.yu.edu]On Behalf Of VERNON SHOUP
Sent: Wednesday, December 08, 1999 1:37 AM
To: Recipients of ABRF List
Cc: VERNON SHOUP
Subject: Light scattering detectors
Importance: Low
We are considering buying an economical light scattering detector for HPLC,
for characterizing protein multimers/aggregates and PEGylated proteins.
Wyatt's low-end instrument is the triple-angle (45, 90, 135 degrees)
MiniDawn. Precision Detectors' low-end instrument is the single-angle (90
degrees) PD2010. More advanced instruments use more angles and/or QELS.
I'd appreciate some feedback from light-scattering detector users. In
particular, for my purposes, does one need multiple-angle detection, or is
just a scattering angle of 90 degrees sufficient? How important is
temperature control? How easy to use are the various manufacturers' software
packages?
Finally, can someone tell me what QELS (Dynamic Scattering) is?
Vernon
Vernon A. Shoup
Regeneron Pharmaceuticals, QC Dept.
Rensselaer, NY 12144
vernon.shoup@regpha.com