Biophysical Characterization – answers to outstanding questions
Earlier this month Mark Pothecary and Paul Clarke (Product Manager and Product Group Manager respectively for Nanoparticle and Molecular Characterization) presented a webinar on a variety of analytical technologies suitable and of interest to biopharmaceutical industry. The recorded event touched on well-established characterization techniques, including light scattering (LALS/RALS/DLS) and size exclusion chromatography (GPC/SEC) and illustrated the benefits of combining complementary information for protein scientists.
This webinar was very well attended, and can still be viewed on the C&EN website here. However, due to the significant number of attendees, not all questions could be answered ‘live’ – so we thought it might be useful to post the Q&A we did not get to during the event. For many issues, expanded details are available on our website, or by contacting us directly. The answers to the questions by the audience were kindly provide by Mark Pothecary, and where appropriate, a few extra links were added here for additional reading.
Q: One slide was about pepsin, where the light scattering method gave 35.2 KDa. I heard before that Light Scattering (LS) measurement have a low limit of around 50 KDa and viscosity measurement can go lower. Is Malvern’s light scattering method more advanced?
A: There are two points to this. To the first, yes, we believe that Malvern’s light scattering instruments do have better sensitivity than other manufacturers. We have no problem measuring smaller proteins such as Lysozyme at around 15kDa. The second part is to say that LS and viscosity have different sensitivities because they are measuring different properties. IV is more sensitive if the molecule is large and open. However it is less sensitive if the molecule is small and compact. So actually for proteins, we find that the sensitivity of LS is actually quite a lot more than IV because they are such tightly packed and compact globular molecules.
Q: Which of these techniques will work with non aqueous systems? All my matrices are organic solvents or mixed organic liquids, I do use GPC and have DLS capability -please comment on the relevance of these techniques for non aqueous matrices.
A: Both GPC and DLS can be used in organic situations. GPC is very commonly used to measure the molecular weight of organically-soluble polymers. We have many other application notes and webinars on these topics so I would suggest having a look at the website. DLS can also be performed in organic solvents although I think this is slightly less common. However, as long as you know the refractive index of the solvent and its viscosity that shouldn’t be a problem.
The one point your question does raise is that you mention matrices. If you are talking about some kind of gel structure, then it would need to be fully dissolved for GPC or broken up for DLS measurements to be successful. If there is no free diffusion, or the sample is not soluble, then neither of these techniques will work.
Q: SLS (Static Light Scattering): is it also done using a cuvette? Can you use it with SEC?
A: Yes, SLS can also be performed in a cuvette, and can be used with SEC. The main point to note is that in a cuvette, you are measuring the signal from all the sample at the same time as it is an ensemble technique. On the other hand SEC separates the molecules in the sample, and resolves the different populations.
Q: What is the low molecular weight limit for DLS and SLS? What is the precision? What is the low molecular weight limit for viscosity measurement? What is the precision?
A: For DLS and SLS, our low molecular weight standard of choice is usually Lysozyme. This is a 14.7 kDa protein with a hydrodynamic radius Rh of about 1.8 nm. This can be reliably measured by both techniques at a concentration of 0.1-0.5 mg/ml. I use the general rule that I think the molecular weight measurement is accurate to within 2-5% if the dn/dc is well known (this is a big if but for most proteins it is consistent.) {Check the Malvern sensitivity calculator for the relevant values for other molecular weights)
For viscosity measurements, IV is actually sensitive to structure so if the molecule is open and loosely folded it will be more sensitive than globular and compact. Most proteins are at the globular/compact end of the scale. You could probably measure down to the Lysozyme range, although it might require a higher concentration (e.g. 5mg/ml 100 ul injection). The precision is probably going to be a bit less at this level though. e.g. 10% RSD.
Q: Can you clarify further how DLS data is transformed into viscosity data.
A: DLS is not converted into viscosity data – the viscosity is measured by a separate detector in the SEC system.
Q: Is it within the scope of this presentation to briefly discuss assymmetrical flow field flow fractionation vs. the content of this presentation for studying aggregates?
A: The simplest way I would differentiate between the 2 techniques would be by the size of the aggregates measured. SEC is good for ‘soluble’ aggregates up to 50 nm or so. Most columns ccan go up to approximately 100 nm. However, anything larger than that would either be excluded or retained. An FFF system would not suffer that limitation so could go up to sizes much larger (as long as they were still abundant enough to be measured). However, an FFF is a bit more limited on the resolution of monomers and oligomers so there is a bit of a trade off there. In the end I think it comes down to exactly what kind of aggregates you are hoping to measure.
Q: Could the SLS method be regarded as having the disadvantage of losing the correlation of sample structure to standard structure? The SLS method seems to only give a molecular weight, while DLS can give some information regarding the sample structure.
A: I think that what you say is true, if you already know the molecular weight of the sample. Then you can see the change in size with DLS. If you only see it with DLS, then you don’t know if the change is due to increasing molecular weight (oligomerisation or aggregation) or because of a change in structure. The same would be true, I think for IV.
Q: What are the optimal sample concentrations (mg/mL) for DLS detection when coupled with SEC? How do these compare to the typical separation and conditions for SEC, specifically related to adequate SEC separations vs. adequate DLS detection?
A: When DLS is connected to an SEC system, the limitation you are putting on it is that you are correlating for short periods of time. In order to get a good signal, you do need to increase the concentration. For a decent measurement of size by DLS in SEC, I would recommend around 2-5 mg/ml of sample for a 100 ul injection. If you are using smaller columns and injection volumes, then you can usually reduce the concentration because the increased resolution will increase your peak heights so you can do a bit better than that, but that’s a good start.
So when compared with DLS in batch you do need a slightly higher concentration. This is because a batch measurement can run the correlation for a longer period of time. For batch DLS, you can easily work at concentrations <1mg/ml. {see for example our sensitivity calculator to predict the lowest protein concentration for DLS for your molecular weight}
For SEC, this depends on the column sizes as mentioned above. Certainly, light scattering is not as sensitive as say UV so when doing UPLC, you do tend to be pushing the limits of DLS. Static light scattering is a bit easier because all you need is the raw light scattering signal, not the correlation function. For SLS you can use concentrations similar to batch DLS (assuming a 100 ul injection). If you have smaller columns, then you could go with the same concentrations but inject smaller amounts.
Light scattering is definitely more comfortable in the analytical column range (30 x 0.8 cm) but with a good setup it is possible to move into UPLC areas.
Q: Is the static light scattering low angle or right angle for the SEC?
A: Most of the data shown in the presentation was Right angle light scattering (RALS, scattering at an angle of 90 degrees). The reason for this is that almost all proteins are isotropic scatterers i.e. they scatter light evenly in all directions. In these situations, we are most interested in getting the best signal-to-noise and for that, RALS is best. For work with proteins, the LALS is usually only useful for large aggregates.
Q: Could you predict how accurately it can measure the MWt of the molecule?
A: I’m not sure here if you are referring to DLS estimates of molecular weight or SLS measurements. I generally work on the rule that SLS is accurate to within 2 – 5 % if the dn/dc is accurately known (which is a fairly big if). Some people would say 10% but I think it is a bit better than that.
If you are talking about estimates of molecular weight from DLS, then it is very unpredictable. If a globular protein, then maybe within 20% but if you are not certain of the structure, then the error could be more.
Q: Does long range interactions affect the data, are these taken into account during measurements of size and viscosity of these proteins?
A: DLS assumes that there are no interactions going on between the molecules when the measurement is made. You are right that this may well be a flawed assumption. If that is the case, then we recommend running the measurement at a series of concentrations and extrapolating the size back to the size at 0 concentration. This is taken to be the ‘true’ Rh value for the sample. The slope of this line is an indicator of the forces between the molecules and is used to calculate a value called Kd. I have attached an application note on this topic that you may be interested in.
If this question was referring to intrinsic viscosity measurement, then generally this is not a problem as by the time the sample reaches the viscometer, the concentration is so low that these effects are negligible.
Q: What are the molecular weight limits for SEC-DLS?
A: I have not made measurements of anything smaller than Lysozyme by SEC-DLS. In order to get good measurements you need a decent concentration of such a small protein (e.g 2 mg/ml, 100 ul injection) but it is certainly possible. For the upper limit, I haven’t found one! These measurements become easier as size increases because of the added signal. I have actually made SEC-DLS measurements of 1 micron particles, so I don’t think there would be a practical upper limit for protein measurements.
For further details:
Press release, summarizing the event and pointing to the just released recording.
Previously
- Improve your SEC-LS data with these 5 simple steps
- Center of SEC Excellence is one year old
- How to obtain reliable info from light scattering