Label-Free Microscale Thermophoresis Discriminates Sites and Affinity of Protein–Ligand Binding

Look, no label! Microscale thermophoresis makes use of the intrinsic fluorescence of proteins to quantify the binding affinities of ligands and discriminate between binding sites. This method is suitable for studying binding interactions of very small amounts of protein in solution. The binding of ligands to iGluR membrane receptors, small-molecule inhibitorss to kinase p38, aptamers to thrombin, and Ca(2+) ions to synaptotagmin was quantified.

To determine the affinity of a ligand L binding to a protein P via label free MST, the total concentration of the ligand c L0 is titrated, while the total concentration of the protein c P0 is kept constant. For the binding event of L to P, the mass action law reads dissociation constant c L : free ligand concentration c P : free protein concentration c LP : concentration of the bound complex c L0 : total concentration of the ligand and c P0 : total concentration of the protein.
Solving for the fraction of occupied binder B yields

Number of repeats and error bars
The number of independent repeats was at least 3 for all measurements except for syt1 and iGluR6 (n=2); error bars show the standard deviation between these independent repeats. For iGluR2 binding glu-azo and the control with the saturated amount of glutamate and with the labeled protein n=1. In these cases the error was estimated by the noise of the fluorescence detection.

Exact error estimations for values given as an upper limit
The fitting procedure assumes a Gaussian, symmetric error distribution. Due to this approximation, the distribution reached negative values in three cases. The corresponding values are given as upper limits of affinity in the manuscript. The exact error estimations are listed in Table S-1.    5 Additional experiments

Aptamer binding to thrombin
Despite their much simpler production and higher stability, DNA aptamers resemble the binding behavior of antibodies. We demonstrate the applicability of label-free MST to monitor the binding of these valuable molecular tools using the example of human α-thrombin (36.7 kDa; 9 Trp residues). This serine protease is part of the human coagulation cascade. In 1992 Bock et al. designed a single stranded 15mer DNA aptamer binding to the fibrinogen recognition exosite. [1] Five years later Tasset et al.
reported on a 29mer binding to the heparin exosite. [2] For our label-free MST analysis of these aptamers, we used a constant thrombin concentration of 200 nM (Fig. S-4). The measured K D of 32±15 nM for the 15mer reproduces the previously reported standard MST result with 5'-fluorescently labeled 15mer (K D =30±19 nM). [3] It is in good agreement with literature values specifying the affinity as 25 to 100 nM based on filter binding assays. [1,2,4] For the 29mer, label-free MST reported an affinity of 133±42 nM. This is much higher than the K D of 0.5 nM, reported by Tasset et al. [2] However, SPR measurements could not confirm the very high affinity of the 29mer either and instead reported a K D in the range of 100 nM. [5] Tang et al. even found an affinity lower than for the 15mer consistent with our results. [6] Apart from the specific aptamers a 15mer dinucleotide mutant was measured. The mutant did not show binding, thus proving specificity.

Figure S-4.
Thrombin interaction with aptamers. Specific 15mer and 29mer DNA aptamers bind to the fibrinogen and heparin exosites of human α-thrombin (c=200 nM). We measured a K D of 32±15 nM for the 15mer and a K D of 133±42 nM for the 29mer and. A 15mer dinucleotide mutant did not bind to thrombin (control).

Ca 2+ -ion-binding to synaptotagmin 1
As previously reported protein binding to ions can be measured with standard MST. [7] This type of binding events is also accessible via label-free MST. We analyzed the interaction of the synaptic vesicle protein synaptotagmin 1 (Syt1) with Ca 2+ . The cytoplasmic Cterminal part of the protein consists of two C2 domains, C2A and C2B. NMR suggests that three and two Ca 2+ bind to C2A and C2B, respectively. Upon binding, the C2 domains mediate membrane translocation. Hence synaptotagmin can act as a neuronal Ca 2+ sensor triggering exocytosis of neurotransmitters into the synaptic cleft. [8] We used Syt1's C2AB fragment in a constant concentration of 1 μM and titrated CaCl 2 . Label-free MST yielded an overall Ca 2+ affinity of 326±26 µM (Fig. S-5). A standard MST control with fluorescently labeled C2AB confirmed this affinity (see Supporting information Fig. S-6). Our results are in good agreement with ITC measurements by Radhakrishnan et al. [9] Assuming quintuple binding, the group extracted K D s of 48, 142 and 3120 µM for C2A and 490 µM for both Ca 2+ binding to C2B. To obtain distinct K D s for the different binding sites, MST measurements of selective Ca 2+ binding mutants of C2AB seem to be very suitable. [10] Titrating MgCl 2 did not result in a change in thermophoresis, demonstrating the specificity of the label free MST analysis (Fig. S-5). Standard MST experiments on cooperative Ca 2+ and phosphatidylinositol 4,5bisphosphate binding to Syt1 have already been performed successfully, suggesting that label free MST could also be used for extensive studies of Syt1 function. [11]

Additional experimental section
Human α-thrombin was purchased from Haematologic Technologies Inc. (Essex Junction, USA). DNA oligonucleotides were synthesized by Metabion (Martinsried, Germany). The sequences of the oligonucleotides, with mutations as small letters, are: 15mer 5'-GGT TGG TGT GGT TGG-3'; 29mer: 5'-AGT CCG TGG TAG GGC AGG TTG GGG TGA CT-3';f 15mer dinucleotide mutant: 5'-GGT TGt TGT GGT TtG-3'. The aptamers were denatured and re-natured prior to the experiments to ensure that they reached their active conformation. The Syt1 C2AB construct from R. norvegicus was cloned into a pET28a expression vector. The purified protein was kindly provided by Geert van den Boogaart.
For the standard MST control Syt1-C2AB was labeled using the Monolith NT Protein Labeling Kit RED according to the supplied protocol.