See also our LS services page
Two different basic light scattering methodologies can be used to characterize proteins:
More details about these methods and their applications are given in three (long) sections below.
Classical light scattering involves measuring the amount of light scattered by a solution at some angle relative to the incident laser beam. For globular proteins smaller than ~500 kDa, the intensity of the scattered light is uniform in all directions, so it is only necessary to measure scattering at a single angle (usually 90 degrees). Theory tells us that the intensity of this scattered light will be proportional to the product of the protein concentration (in mg/ml) times its molar mass.
For higher masses or extended proteins (rod-like or unfolded) the scattering varies significantly with angle. By measuring the scattering at additional angles ("multi-angle light scattering", or MALS), direct absolute measurements of masses up into the MDa range can be made, and the "r.m.s. radius" (a measure of geometric size) can also be determined.
This technique is generally best used on-line in conjunction with size-exclusion chromatography (SEC-MALS), as shown in this diagram.
Since the signal from the light-scattering detector is directly proportional to the molar mass of the protein times the concentration (mg/ml), by combining this signal with that from a concentration detector (refractive index or absorbance) it is possible to measure the molar mass of each peak coming off the column.
Unlike conventional SEC methods, these molar masses from light scattering are independent of the elution volume. Thus this technique can be used with "sticky" proteins that elute unusually late as a result of their interactions with the column matrix, and also with highly elongated proteins which elute unusually early for their molar mass. The molar masses derived by this technique are generally accurate to 3% or better.
The amount of protein required for this type of analysis is inversely related to its molar mass. For a 20 kDa protein it is helpful to have ~100 micrograms (but often one can get by with much less). In some cases it is useful to inject quite large amounts of protein in order to get better signal/noise for the very minor peaks.
As one example, below are results from some studies of an antibody sample by Alliance Protein Laboratories. The individual data points (+) show the molecular weight at each point in the chromatogram, while the solid line shows the overall elution profile as detected by the refractive index (RI) detector. These data, as expected, proved that the peak eluting around 13.4 ml is a dimer. However, the shoulder of material eluting before the dimer was unexpectedly found to also contain dimer, rather than trimer or tetramer as had been assigned based on elution position. Based on these data, as well as sedimentation velocity studies, A.P.L. was able to show that there are two different conformations of dimer in this material.
In contrast, studies of a different antibody (below) detected very small amounts (2.6% total) of dimer and trimer.
The amount of aggregate eluting in the region from ~12.7 to 13.6 ml (<.1%) is too small for reliable mass estimation by the traditional approach using the RI signal as the concentration reference, but A.P.L.'s own custom "oligomer hunting" software used here gives higher signal/noise and quickly reveals how many monomers are present in each aggregate peak by presenting the data as the mass ratio relative to the main peak.
This light scattering + SEC approach is also excellent for studying whether different manufacturing procedures and different lots give equivalent amounts and size distributions of aggregates. As shown below, a different lot of this same monoclonal gave a much higher total amount of aggregate and a distribution that extended to much higher mass (at least dodecamer).
This method is also very useful for characterizing conjugated molecules, like PEGylated proteins or nucleic acids, or glycoproteins. It can identify whether the conjugation has caused the original molecule to aggregate, and also measure the extent of conjugation. APL scientists pioneered approaches for analyzing the light scattering from conjugated molecules (see papers available from our Further Reading page), and algorithms based on their work are now incorporated into the data analysis software from Wyatt Technology.
Alliance Protein Laboratories has a Wyatt Technology miniDAWN™ TREOS MALS light scattering detector used together with both a refractive index detector (Wyatt Optilab rEX) and a UV absorbance detector [photo of SEC-MALS setup here].
APL has recently purchased a Wyatt Technology Calypso accessory for our MALS detector so we can offer Composition-Gradient MALS (CG-MALS) analysis [picture of CG-MALS setup here]. As shown schematically below, CG-MALS combines computer-controlled syringes and a mixing system that flows the mixed solutions into the MALS detector and then on into a concentration detector (UV or RI).
This technology measures the weight average molar mass of solutions as a function of composition (total concentration, and as a function of mixing ratio when studying interactions between two different macromolecules). In many ways CG-MALS is similar to sedimentation equilibrium (SE-AUC), but the experiments are more rapid, and it is possible to obtain data at very high concentrations (up to ~100 mg/mL). CG-MALS can also be used for automated measurements of second virial coefficients (to assess colloidal stability under different solution conditions).
Introduction to CG-MALS: example data for BSA
The graph below shows a CG-MALS experiment where a solution of BSA at ~8 mg/mL in DPBS was diluted to zero concentration in six equal steps. We see an initial build-up of the scattering and RI signals to a plateau level as the flow system is saturated with the BSA stock, and then a series of six downward stair-steps as that stock is mixed with increasing amounts of buffer.
The drop in scattering intensity with each drop in concentration can then used to calculate an apparent molar mass at that concentration, as shown in the graph below.
The slight drop in apparent mass as the concentration increases is due to repulsive solution non-ideality (“molecular crowding”) effects. Fitting to a straight line gives the expected 66 kDa mass at zero concentration and also the second virial coefficient.
Using CG-MALS to characterize strong reversible association of a monoclonal antibody
The graph below shows the raw CG-MALS data for a monoclonal antibody which strongly reversibly associates at concentrations of a few mg/mL. This sequence shows a 10-step downward dilution from a starting concentration of ~1.5 mg/mL (with UV concentration detection at 297 nm in this case).
Clearly the scattering and concentration traces are much more divergent than for the BSA data above (and in the opposite direction). Here we see that the initial downward LS steps are much larger than the later ones, which means the effective molar mass is falling as the concentration falls. Another important observation is that the light scattering reaches a plateau quickly after each dilution, with no evidence of slow dissociation of reversible oligomers.
Two additional similar titrations were done starting from ~0.5 and ~0.2 mg/mL, giving 30 total concentration points. The graph below summarizes all those data as an apparent molar mass vs. concentration plot. This plot shows that near 1.5 mg/mL the weight-average mass exceeds that of dimer, and hence oligomers at least as large as trimer must be present.
Attempts were made to fit the mAb CG-MALS data to both non-specific
indefinite association (isosdesmic) models and to specific assembly
stoichiometries. A monomer « dimer
« tetramer model gives a reasonably good fit
(summarized in the graphs below) and returns dimer ®
monomer and tetramer ® dimer dissociation constants
of 15 μM (4.5 mg/mL) and 3.5 μM
(2.1 mg/mL), respectively.
Other downloadable CG-MALS examples
The above monoclonal antibody CG-MALS data are compared to results from sedimentation velocity (SV-AUC) for this mAb in this poster by John Philo from the 2013 Higher Order Structure meeting.
This invited talk by John Philo at the 2014 Higher Order Structure meeting gives an overview of applying AUC and CG-MALS to characterize reversible association of proteins and peptides, including a discussion of the difficulties arising at concentrations above 50 mg/mL.
In dynamic light scattering one measures the time dependence of the light scattered from a very small region of solution, over a time range from tenths of a microsecond to milliseconds. These fluctuations in the intensity of the scattered light are related to the rate of diffusion of molecules in and out of the region being studied (Brownian motion), and the data can be analyzed to directly give the diffusion coefficients of the particles doing the scattering. When multiple species are present, a distribution of diffusion coefficients is seen.
Traditionally, rather than presenting the data in terms of diffusion coefficients, the data are processed to give the "size" of the particles (radius or diameter). The relation between diffusion and particle size is based on theoretical relationships for the Brownian motion of spherical particles, originally derived by Einstein. The "hydrodynamic diameter" or "Stokes radius", Rh, derived from this method is the size of a spherical particle that would have a diffusion coefficient equal to that of the protein, and the data is commonly presented as the fraction of particles as a function of their diameter.
Most proteins and peptides are certainly not spherical, and their apparent hydrodynamic size depends on their shape (conformation) as well as their molar mass. Further, their diffusion is also affected by water molecules which are bound to them or entrapped. Therefore, this hydrodynamic size can differ significantly from the true physical size (e.g. that seen by electron microscopy or x-ray crystallography), and this size is generally not a reliable measure of molar mass.
One important application for DLS is measuring the hydrodynamic size of molecules that have been PEGylated to prolong serum lifetime and/or to 'hide' the molecule from the immune system or specific receptors. In such cases the increase in hydrodynamic size can be an important measure of how effective the conjugation will be in improving the product efficacy or safety. DLS directly and accurately measures the true hydrodynamic size, unlike indirect methods such as SEC that rely on standard molecules and questionable assumptions.
Utility of DLS for studying aggregation
While dynamic scattering is, in principle, capable of distinguishing whether a protein or peptide is a monomer or dimer (their radii would be measurably different), it cannot resolve monomer from small oligomers, and cannot quantitate fractions of small oligomers. In general two different species need to differ in hydrodynamic radius by a factor of 2 or more in order for them to be resolved as separate peaks. A factor of 2 in radius corresponds to roughly a factor of 8 in molar mass! Thus typically monomer, dimer, trimer,... octamer will all be merged into one average peak. Consequently DLS is much less used for analyzing small oligomers than is SEC-MALS or sedimentation velocity.
The three key strengths of dynamic scattering are:
Below is an example of a sample with a main component at ~2.2 nm and two clear, well-resolved aggregate peaks at 6.6 and 92 nm. Although the second aggregate peak represents 13.3% of the total scattered light, because that intensity is proportional to molar mass, and mass increases as (radius)3, the fraction by weight represented by this aggregate is only ~0.015% of the total!
A common problem with proteins and peptides is the appearance of sub-visible or visible particles ("snow" or "floaters") over time. Generally these particles grow from trace amounts of much smaller precursor aggregates. Below are some DLS data for "bad" lot of small peptide that exhibited problems with visible particulates (after centrifugation to remove large particles). The scattering intensity is dominated by a peak at ~100 nm radius (only 0.002% by weight!) which appears to a precursor to the visible particles.
Thus DLS can be a good way (and often the only way) to detect such precursors long before visible particles are present (which may take months). Thus this technique can provide a quick assay to help track down and prevent formation of visible particulates, and we have successfully solved such problems for a number of clients using this approach.
Another advantage of dynamic scattering is the ability to study samples directly in their formulation buffers or at high protein concentrations (50 mg/ml or more). It should be noted, however, that the interpretation of data from high concentration samples (and especially the values of the apparent hydrodynamic radius) can be quite difficult due to solution non-ideality ("molecular crowding") effects. Nonetheless, as in the data below for a protein (an IgG) at roughly 50 mg/ml, the fact that the sample is not homogeneous and contains some aggregates at ~20-40 nm radius is unambiguous.
Dynamic scattering can be quite useful for assessing aggregate formation over time, and directly comparing rates of degradation for different formulations. Accelerated stability studies can be carried out by monitoring a single sample in situ (made possible by the Peltier temperature control in our instrument) or by periodic sampling of multiple samples held at elevated temperature.
Probably the primary drawback of dynamic scattering is that it is often difficult to accurately quantitate the amount (weight fraction) of any aggregates which may be present. Nonetheless it can be a very good technique for relative comparisons, such as indicating which formulation, sample treatment, or purification process produces more aggregates.
Alliance Protein Laboratories has a Wyatt Technology Nanostar instrument for DLS studies (new in 2013), a state-of-the-art system which uses very small sample volumes (~1 microliter) and which is equipped with Peltier temperature control. [photo in lab here]
See also our LS services page
All text, images, and downloadable documents are © copyright 2015, Alliance Protein Laboratories Inc.