Exploring Weak Ligand–Protein Interactions by Long-Lived NMR States: Improved Contrast in Fragment-Based Drug Screening

Ligands that have an affinity for protein targets can be screened very effectively by exploiting favorable properties of long-lived states (LLS) in NMR spectroscopy. In this work, we describe the use of LLS for competitive binding experiments to measure accurate dissociation constants of fragments that bind weakly to the ATP binding site of the N-terminal ATPase domain of heat shock protein 90 (Hsp90), a therapeutic target for cancer treatment. The LLS approach allows one to characterize ligands with an exceptionally wide range of affinities, since it can be used for ligand concentrations [L] that are several orders of magnitude smaller than the dissociation constants KD. This property makes the LLS method particularly attractive for the initial steps of fragment-based drug screening, where small molecular fragments that bind weakly to a target protein must be identified, which is a difficult task for many other biophysical methods.


Experimental procedures
The direct titration of ADP (ligand I) against Hsp90 was performed by adding 2-5 μL aliquots of ADP stock solution to 400 μL of a 10 μM solution of Hsp90. Titrations of ligands II and IV were carried out with the same procedure. where pulses with flip angles of π/2, π/4 (45°) and π are indicated by white, grey and black rectangles, respectively. For a system comprising two non-equivalents spins I = ½ and S = ½ with a chemical shift difference ΔνIS = (ΩI -Ωs)/2π and a scalar coupling constant JIS, the intervals τ1 and τ2 must be set to: At the beginning of the experiment, a π/2 pulse followed by a pulsed field gradient along the z axis were applied to suppress the magnetization. A weak water presaturation pulse lasting 3 s was applied in the subsequent recovery delay. During the sustaining delay 1.2 s < τm < 3.5 s (chosen depending on the expected the constant TLLS = 1/RLLS), continuous-wave radio-frequency (rf) irradiation was applied, with the carrier set half-way between the two chemical shifts, Ωcarrier = (ΩI+Ωs)/2, in order to make the spins magnetically equivalent in the sense of average Hamiltonian theory. Two pulse field gradients were applied during the two τ2 delays in order to dephase double quantum coherences. The acquisition time was 2.2 s. The phase cycle was Φ1 = (x,-x), Φ2 = (y,y,-y,-y), Φ3 = (y) and Φreceiver = (x,x,-x,x). The number of scans was chosen according to the concentration. For example, during screening experiments with 500 µM of spy ligand, 64 scans were accumulated, thus requiring 10 minutes per sample. The rates RLLS were obtained from the ratio of the signal intensities observed using two different sustaining delays a and b, repeating each of them twice in order to compare four pairs of signal intensities Ia(a) -Ib(a): Typically, we used a = 0.5 s and b in the vicinity of the estimated value of TLLS = 1/RLLS.

LLS titration of a spy molecule
The LLS titration of the spy molecule without competitor was carried out by addition of aliquots of a stock solution of 150 mM vanillic acid diethylamide in DMSO to 400 μL of a 10 μM solution of Hsp90 protein in the buffer described above. The contrast CLLS was determined after sustaining the LLS for a = 0.5 s and 1.0 < b < 2.5 s.
By fitting the relaxation rates to equation 5, one can determine the dissociation constant . One should perform two kinds of titration experiments: Direct titration. If the spy ligand that carries the LLS is titrated against the protein in the absence of any competitors, one obtains = , , i.e., is the true value of the dissociation constant of the spy ligand.
Competition titration. If the spy molecule is titrated against the protein in the presence of a competitor, one initially obtains = , the apparent dissociation constant of the spy ligand. Note that , > , . The latter can be extracted by LLS titration without competitor, or by using non-NMR methods such as ITC.
One may calculate the dissociation constant of the competitor with the following relationship [2]: where is the dissociation constant of the competitor and [ ] is the concentration of the competitor.
The errors (root-mean-square deviations) were estimated from the four pairs of signal intensities obtained using two different sustaining delays a and b and repeating each experiment twice.