If your protein is an enzyme, and if you are looking for inhibitors, and if you have an assay for the enzyme activity, that is an easy way to test the compounds.

Hi Mohamad, the cheapest and fastest way I can think of is to immobilize your purified protein and pass the small MW compounds through the resin to see if they bind.

Hi there,

If you have the device in hands I think thermal shift assay is definitely the easiest as you don't have to separate the complex from the free partners: ligand binding leading to a more stable protein. For more info:

Article Current Protocols in Protein Science

@ Steingrimur Stefansson Thanks, could you please reference a protocol to do that? The proteins need to be tagged in this approach?

Hi Mohamad, I don't have a specific protocol for measuring small molecular weight compounds binding to immobilized protein because it depends on the nature of the compounds and how you measure them.

Immobilizing proteins on a resin is easy and you dont need a tag to do it.

http://www.bio-rad.com/webroot/web/pdf/lsr/literature/LIT156.pdf

I also think the thermal shift assay is the simplest, especially if you want to test multiple compounds in parallel.

If your protein is an enzyme, and if you are looking for inhibitors, and if you have an assay for the enzyme activity, that is an easy way to test the compounds.

You do not mention what instruments you have available to you, but 20 compounds would be a simple project for both MST and SPR, but these are expensive instruments.

Perhaps the simplest way is the use of spin columns: Fill a 1 ml tuberculin syringe with 0.8 mL of a gel filtration medium that excludes your protein, but includes the ligand (G10 should be fine for most purposes). Porous PE sheets (e.g., Bel-Art H13638-5116) are convenient to cut filters to be placed at the bottom of the syringe. The gel should be swollen in the same buffer that is used for the assay, but w/o ligand. Let excess fluid drain and then place the syringe in a centrifugation tube, so that it hangs from the "wings" at the top. Add a mixture of protein and ligand on top of the gel, say 50 to 100 µl. Spin briefly in a clinical centrifuge, so that the protein with bound ligand is in the eluate (void volume) at the bottom of the tube. Unbound ligand will remain on the column. Important is that the time the protein needs to pass through the gel is short compared to koff. Then determine the concentration of protein and ligand in the eluate. Proper gel filtration columns can also be used, they give more precise values, but take a lot more time (DOI=10.1016/0006-3002(62)90124-5).

Ultrafiltration in spin filters is another method (DOI=10.1006/abio.1998.2854). The cutoff of the membrane should be chosen so that the ligand can pass freely, but the protein is excluded. Spin long enough so that about 90% of the fluid passes through the membrane. Measure the difference of total [L] between flow-through (free ligand) and the retentate (free + bound ligand). You also must determine the protein concentration of the retentate.

If you have access to an ultracentrifuge, you can also use either Chanutin's method (http://www.jbc.org/content/143/3/753.full.pdf) or that of Draper & von Hippel (DOI=10.1021/bi00572a003).

Bjarte Aarmo Lund No matter! I want to know what is the cheapest approach. If a method is cheap the equipment should be popular in different lab so finding the equipment is not very important in my question.

Engelbert Buxbaum Thanks for your complete reply! Do you have any estimation about the cost of these approaches? How much purified proteins and ligand roughly needed to run the experiments?

Adam B Shapiro So thermal shift assay just works for enzyme? Right?

Protein is a limiting factor in binding studies, you need much more (by several orders of magnitude) than for enzyme kinetics, as you have no amplification. The cost in protein depends in part on the sensitivity of your assay for ligand. However, precision of the assay is greatest when [P]t is near (that is, within an order of magnitude of) Kd. Remember that you are interested in the difference between the ligand concentration in fractions with and w/o the protein, and that small differences between large numbers are notoriously error prone.

Assume that your [P] is equal to Kd, then you need at least 12 different [L], evenly spaced between 0.1 and at least 5 Kd (better 10 Kd, if technically feasible). Recall that 5 Kd corresponds to θ = 83% and 10 Kd to θ = 91%, asuming hyperbolic binding curves. Thus, the higher [L] gets, the better is the estimate for the number of binding sites. Low [L], on the other hand, improve the estimate of Kd.

You want to do the assay in triplicate, and you want to repeat it with 3 or 4 independent preparations of your protein. This will give you an idea of the reproducibility of your assay (intra-day and inter-day). So you need ~ 150 assays to reliably determine one Kd value. If you need to determine Kd as function of some other factor (e.g., inhibitor concentration), multiply this number accordingly. In such a case, however, you may seriously consider sparse matrix approaches to experimental design (i.e., a Taguchi orthogonal matrix) to drastically reduce the number of experiments required.

Thermal shift can be used for any protein in solution, not just enzymes. I meant that you can use an enzyme activity assay to test your compounds for enzyme inhibition, which would be an indication that the compounds bind to the enzyme, and also give a measurement of how tightly.

The best way to do the binding experiment is the method that is available with the materials and equipment that you have. If you have expertise in NMR, you could do a ligand-observe NMR experiment, which doesn't need much protein. If you have access to an isothermal titration calorimeter and a lot of protein, you can do an ITC measurement. The methods mentioned by Engelbert Buxbaum are pretty easy and fast, especially using miniature centrifugal ultrafilters, as long as you have some way of measuring the compound concentration, and it passes freely through the filters rather than sticking to them. If you have fluorescence spectroscopy experience, you might be able to use the intrinsic protein fluorescence to measure compound binding. You can also consider the Hummel-Dreyer method, which is based on chromatography. There are many possibilities.

One of the best methods, in my opinion, to demonstrate binding to a protein by weakly binding small molecules is protein-observe 2-D NMR. This sophisticated method required specially prepared protein, and a high-end NMR instrument, but it gives highly definitive results and can yield a Kd for the compound. SPR, already mentioned by Bjarte Aarmo Lund, can also give a Kd measurement, but also requires expensive equipment and special expertise.

Adam B Shapiro Thanks a lot for your complete reply!

How about doing a solid phase binding assay?

Spectroscopy equipment may be available at your institute, depending on the nature of the ligands, either the protein or the ligand may produce a signal upon binding, that is not observed for the free ligand and the free protein.

Using appropriate blanks, you can do serial additions of the ligand to your protein. In cases where one known ligand does produce a signal, but your molecules of interest do not, you can do competition assay. Place a known ligand and the protein in the cell at fixed concentrations and perform serial additions of the ligand to be tested. A reduction in the signal of the complex indicates displacement and can be used to calculate an apparent Kd, if proper controls are used, and a correction for the dilution effect is done.

Even better is ITC (isothermal titration calorimetry), here your ligand binds to the protein and either absorbs heat or emits heat. In any case, the equipment compensates the heat by cooling or heating to keep constant temperature. The temperature trace is integrated to calculate the heat introduced or withdrawn by the calorimeter. Using serial additions of ligand and after correction for dilution effects, the Kd can be calculated.

Thermal shift (mentioned above) is a variant of this principle, based on the difference in stability of the Complex vs. the free protein. Here, the energy involved in the ligand-protein has to be large enough as compared to the protein stability, and the data can only be used to obtains an apparent Kd approximate estimation.

All of these techniques require a large amount of protein, except for thermal shift.

Best wishes,

Rogelio

Abir Chakraborty solid phase binding assays can give you Kd values in many cases using bound protein and soluble ligand. ELISA's will not give you a Kd.

In my personal experience, the fastest, simplest, and cheapest method to measure binding of compounds to a target protein (or nucleic acid) is by a competitive displacement assay using fluorescence intensity or polarization in microplates, using a fluorescence plate reader. The method requires a fluorescently labeled ligand that occupies the binding site of interest and exhibits a difference in fluorescence intensity or polarization depending upon whether it is bound or free. The method also requires the availability of the specialized plate reader and a small supply of the appropriate black microplates. The assay, once developed, can be carried out in only a few minutes and, if miniaturized to a 384-well format, can use a very small amount of protein.

Cheap?

Not counting the cost of the plate reader.

Thanks all to share your idea! That is really helpful!

Hi Mohamad,

We do affinity selection mass spectrometry based ligand screening on a daily basis in the lab. Refer to the following manuscript and in your case, I would recommend the pulsed ultrafiltration with a cut off filter column.

M. D. Rush, E. M. Walker, G. Prehna, T. Burton, R. B. van Breemen. Development of a Magnetic Microbead Affinity Selection Screen (MagMASS) Using Mass Spectrometry for Ligands to the Retinoid X Receptor-α. J. Am. Soc. Mass. Spectrom. 2017, 28, 479.

@Ruth Muchiri

Thanks for your reply. Do you have a rough idea about the price of a magnetic microbead which is used in this method to immobilize the target protein?

For the assay reported with the RXR, the amylose magnetic beads were about $250, minus the other reagents and supplies. This assay wold require that you have a maltose binding protein tag on your protein. The other magnetic beads were the NHS, which would require immobilization step (this one was a little cheaper - $150 for 10 mg/mL beads) and as long as the protein active site is not altered by the immobilization step, you can go for it (may take you longer to have the assay going though).

If possible, probably consider the molecular filter based screening option, if you do not mind spending a little more time on the assay and save some cash.

One more thing, you do not have to test your compounds singly, you can divide them into 4 (5 compound per assay) and in total you would need 10 assays (4 assays with your compounds and intact protein, 1 positive control with a known ligand for your target, and 1 negative control (with a non binder, if known), and then 4 assays with the denatured protein). And you can do this in triplicates within a day, including the analysis.

I am seeking collaborators for a target protein project. This is a quick and simple novel assay for measuring protein-ligand interactions. Would you be willing to provide purified proteins and native ligands for my project? If so, send me a message.

Thank you Mr. Adam B Shapiro for your valuable answer.