Q&A with Dr Devenish – common questions about MDS analysis

Published on October 10th, 2019

In a quick Q&A session Dr Sean Devenish answers the most commonly asked questions about MDS and its application to protein binding studies.


Question: Is this method able to assess Protein–DNA binding?

Answer: Yes, there are two ways of doing this. The aptamer example I used earlier was one where the DNA was labelled, we were looking at the DNA binding to the protein. However, you can also flip that around and look at protein binding to DNA. The key constraint there is that if you label the protein, then you need to have a larger piece of DNA so that we can measure a good size change. So, using long oligomers or using excised fragments of plasmids, or DNA structures of around 100-200 base pairs is optimal. In fact, we are looking at transcription factors in the lab at the moment. But to summarize, if the sample is fluorescent at 488nm wavelength and the species or complex is between 0.5 nm-20 nm then anything would work; sugars, lipids, proteins, DNA, small molecules etc.

Question: Is it possible to measure membrane protein binding using this approach?

Answer: It is yes. This is an area where we have had a lot of interest. The advantage we have is that because we are in solution, membrane proteins can be presented in a more natural context like in the context of liposomes, lipid rafts or nanodiscs for example. Because we don’t need to fix the protein on the surface there’s no risk of conformational effects or any of the disadvantages you’d see with methods where fixing is required. The workflow would be to take your membrane protein of interest and titrate it against a labelled binding protein, then measure the sizes and get the binding curves as shown in the presentation.

Question: Can affinity be measured in crude biological fluids?

Answer: It can, the first example in the presentation touched on this when we used cell culture medium. We have a partner who is a clinician at a hospital here in Cambridge who is measuring binding affinities in blood serum and is having a lot of success in that endeavor even with minimal dilution of blood serum.

Question: Why is this approach suited to intrinsically disordered proteins (IDPs)?

Answer: So, intrinsically disordered proteins (IDP) are a fascinating class of proteins in that they function in a folded form but that they often exist in an unfolded form. The transition they make from folded to unfolded is functionally relevant. They play a role in a lot of binding interactions too so they’re really important and are becoming of interest to pharmaceutical companies. The problem with IDPs are that, in terms of energy, they live on the cusp of between folded and unfolded, they live just on the unfolded side. So, if you start changing their environment too much then they start to fold or partially fold, so the interaction you measure might not be representative to the one that takes in a biological system.

This is always the risk you take when you use surfaces, if you attach an IDP to a surface or a matrix element like a gel, then that protein may adopt a fold of some sort. Because our measurement that takes place completely in solution, the only species that an IDP would meet would be its binding partner or other IDPs. By working with proteins in their natural state the Fluidity One-W can achieve more representative results.

Question: What do I need to do to prepare my samples before using this technique?

Answer: It’s pretty straight forward. One species needs to be labeled, there are several kits that can be used but we are looking to put one together ourselves in the future. It is preferable to label the smaller of the two species because it works better that way. But you can use most types of labels including but not limited to: FITC, AlexaFluor, NHS, Maleimide chemistry and GFP. If you have used a chemical label it is necessary to remove the free dye otherwise it brings down the diffusional size value of the protein. Once you have your labelled protein the titration is all that’s required.

Question: What is the data output from the instrument?

Answer: The instrument will provide a KD measurement on screen and will plot the titration series as you run the experiment which can guide you so that you know which measurements you might want to repeat or when you have enough data points to finish the titration. We know that researchers also like to go deeper so can also extract the size and intensity of each data point, which you can take off and analyze on your software of choice. The machine can also report the raw data on the diffused and undiffused sizes that are measured which can be used to determine the affinities. All of this information is easy to download from the instrument using a USB flash drive.

Question: How does this approach compare to fluorescence polarization (FP)?

Answer: FP is a similar technique, but one of the key differences is that we can work with species that are closer together in terms of size.

FP works by taking a labelled molecule and exciting it with polarized light and the polarity of the emitted light is measured. As the molecule tumbles the polarity of the emitted light is scrambled, so that polarized light reports on the tumbling rate of the protein of interest. So, if the labelled molecule is bound to another protein its tumbling rate will decrease. The main limitation of FP is that because it’s based on rotational motion being measured instead of a translational motion, there’s a requirement for a significant size change. That means the species being measured by FP must be quite different and is limited to measuring species of 10 kDa or less. We can work with far bigger species.

In addition, FP does not give absolute size information about the labeled species, so it will only be able to provide the KD but not the size of or stoichiometry of an interaction which the One-W does report.

Question: Can this technique be performed without prior characterization of the protein?

Answer: Absolutely, the initial calculations you make with little to no concentration of binding partner act as the characterization of the protein anyway. We do provide a calculator on our website which can give you an estimate of the hydrodynamic radius based on the known molecular weight of the protein and whether it is folded or unfolded. So, you can use that as a guide to see whether the hydrodynamic radius you are seeing corresponds to the species you have in solution. You can start to see if your protein is binding to the unlabeled partner you predict in this way too.

Question: Could this technique be used in competition assays or be used to compare two different binding partners?

Answer: Certainly, we are already doing competition assays in the lab to measure affinities where you do have a label but the interaction itself doesn’t involve the label. Essentially, we are displacing a labeled binder with an unlabeled species and determining the binding affinity. In some instances, a competition assay is a preferable way of operating the instrument. It can be used to compare two different binding partners; this can be done in separate experiments by labeling each of the binding partners separately and then titrating them against a common partner.

Question: You have mentioned that this technology can measure protein binding to lipids, how is this feasible in solution? Will using detergents in order to enable solubilization of lipids interfere with diffusion across barriers?

Answer: To be clear, there are no barriers in the diffusion channel. Because there is laminar flow between the sample and the auxiliary streams there is no mixing of solutions. The ideal approach when working with lipids is if they can be presented as vesicles or micelles. If the lipid is soluble then it will be very small so you would have to label the protein. The best approach is to take a lipid vesicle and a labelled protein and titrate the vesicles against the protein to measure the binding affinity for that interaction.