Some sulfonamide drugs inhibit ATPase activity of heat shock protein 90: investigation by docking simulation and experimental validation

Abstract

Eight selected sulfonamide drugs were investigated as inhibitors of heat shock protein 90 (Hsp90). The investigation included simulated docking experiments to fit the selected compounds within the binding pocket of Hsp90. The selected molecules were found to readily fit within the ATP-binding pocket of Hsp90 in low-energy poses. The sulfonamides torsemide, sulfathiazole, and sulfadiazine were found to inhibit the ATPase activity of Hsp90 with IC₅₀ values of 1.0, 2.6, and 1.5 μM, respectively. Our results suggest that these well-established sulfonamides can be good leads for subsequent optimization into potent Hsp90 inhibitors.

Keywords::

• Hsp90α
• sulfonamides
• docking
• pK_a

Introduction

Heat shock protein 90 (Hsp90) belongs to a family of molecular chaperones that play a pivotal role in the conformational maturation, stability, and function of protein substrates within the cell. The ATPase activity of Hsp90 provides the necessary energy required for refolding of denatured cellular proteins.¹

Amongst the client proteins of Hsp90 are many oncogenes essential for the survival, proliferation, invasion, metastasis, and angiogenesis of tumors.² In fact, several oncogenic proteins have been shown to be dependent upon Hsp90 for conformational activation, including: telomerase, Her2 (erbB2), Raf-1, focal adhesion kinase, and the steroid hormone receptors.³

The validity of Hsp90 as anticancer target for drug discovery^4–16 was further established by emerging clinical and preclinical trials employing the potent Hsp90 inhibitor 17-allylamino-17-desmethoxygeldanamycin^17–27 and the natural Hsp90 inhibitors geldanamycin,^28–31 radicicol,³² and other small molecules, for example, purines³³ and pyrazoles.³⁴

However, despite the high cellular activity and clinical progression of 17-allylamino-17-desmethoxygeldanamycin, it has several limitations, for example, poor solubility, hepatotoxicity, and extensive metabolism.^17–27 These issues have led to significant efforts to identify novel small molecule inhibitors of Hsp90.^35–37

The recent discovery that certain hydroxynaphthalene-arylsulfonamide derivatives possess micromolar inhibitory actions against Hsp90^38,39 combined with the moderate anticancer properties reported for antibacterial sulfonamides and their selective accumulation in cancerous cells⁴⁰ prompted us to assess the inhibitory profiles of eight clinically established sulfonamides against the ATPase activity of Hsp90, namely, torsemide, sulfathiazole, sulfadiazine, sulfamethizole, sulfisoxazole, sulfadoxine, sulfaguanide, and sulfacetamide (Figure 1).⁴¹ The excellent safety profiles of the selected compounds should render them superior leads for subsequent optimization if found active. Docking studies into the ATP-binding site of Hsp90 were used to explain the inhibitory bioactivities of active compounds.

Figure 1.  Chemical structures of tested sulfonamides.

The selected compounds were bioassayed against recombinant human Hsp90α (BioQuote, York, UK) employing malachite green-based detection of free phosphate released by the ATPase action of Hsp90^42–44 using geldanamycin as positive control to standardize our experimental setup. We believe malachite green assay is more appropriate to measure the inhibitory effects on ATPase activity of Hsp90 compared with other more biologically intensive methods, for example, western blotting, since these rely on living cell lines that might show problems related to compound permeability across cellular membranes and metabolism, which may not necessarily reflect real Hsp90 inhibition.³⁹

Materials and methods

All of the chemicals used in these experiments were of reagent grade and obtained from commercial suppliers. Torsemide, sulfadiazine, sulfathiazole, sulfamethizole, sulfisoxazole, sulfadoxine, sulfaguanide, and sulfacetamide were purchased from Sigma-Aldrich (Munich, Germany), ATP 100x solution, geldanamycin, and Hsp90α were purchased from BioQuote (UK); Quantichrom ATPase/GTPase assay kit was purchased from BioAssay Systems (Hayward, CA).

Docking

The binding site was generated from the cocrystallized ligand (GMD275) within Hsp90α protein (PDB code: 1YET). All eight ligands were docked employing the following docking configuration: (i) number of Monte Carlo search trials 25000, search step for torsions with polar hydrogens = 15°. (ii) The Root Mean Square Difference (RMS) threshold for ligand-to-binding site shape match was set to 2.5 employing a maximum of 1.0 binding site partitions and 1.0 site partition seed. (iii) The interaction energies were assessed employing Consistent Force Field (CFF) force field with a nonbonded cutoff distance of 10.0 Å and distance-dependent dielectric. An energy grid extending 5.0 Å from the binding site was implemented. (iv) Rigid body ligand minimization parameters were: 40 iterations of steepest descend (SD) minimization followed by 80 Broyden–Fletcher–Goldfarb–Shanno (BFGS) iterations applied to every successful orientation of the docked ligand. (v) A maximum of 10 diverse docked conformations/poses of optimal interaction energies were saved. The similarity threshold was set to a DockScore of 20 kcal/mol and an RMS value of 1.5 Å. (vi) The saved conformers/poses were further energy-minimized within the binding site for a maximum of 200 rigid-body iterations. The resulting docked poses were scored employing consensus scoring based on PLP1, PLP2, ligscore1, ligscore2, PMF, and JAIN.^45–51 The optimal docked poses of torsemide, sulfadiazine, sulfadoxine, and sulfaguanide achieved full consensus score from all six scoring functions, while the remaining compounds failed to do so. Therefore, sulfacetamide, sulfamethizole, sulfisoxazole, and sulfathiazole were superimposed against the docked poses of sulfadiazine and sulfadoxine. Subsequently, they were docked via rigid body docking employing the following settings: (ii) The RMS threshold for ligand-to-binding site shape match was set to 2.5 employing a maximum of 1.0 binding site partitions and 1.0 site partition seed. (iii) Interaction energies were assessed employing CFF force field with a nonbonded cutoff distance of 10.0 Å and distance-dependent dielectric. An energy grid extending 3.0 Å from the binding site was implemented. (iv) Rigid body ligand minimization parameters were: 10 iterations of SD minimization followed by 20 BFGS iterations applied to every successful orientation of the docked ligand. (vi) The saved poses were further energy-minimized within the binding site for a maximum of 100 rigid-body iterations.

Measurement of Hsp90 inhibition

The bioassay was performed as reported earlier.^52,53 In brief, tested compounds were dissolved as dimethyl sulfoxide (DMSO) stock solutions (0.2 M). Bioassays were performed by mixing Hsp90α solution (6 µL, 25 µg/mL in assay buffer), 24 µL assay buffer, and 5 µL of the particular tested compounds to yield final inhibitor concentrations of 100, 10, 1, and 0.1 µM per well. The mixtures were incubated for 30 min at 37°C in ELISA plate shaker, and then ATP solutions (5 µL, 4 mM in assay buffer) were added to each mixture. Blank was prepared as earlier except 5 µL of 2% DMSO in distilled water (v/v%) was used instead of inhibitor solution. The mixtures were equilibrated to 37°C and incubated for 24 h. The enzymatic reaction was terminated by the addition of 80 µL malachite green ammonium molybdate–Tween-20 solution in 0.27 M H₂SO₄ and 10 µL of 34% sodium citrate. Color was allowed to develop at room temperature for 30 min, and sample absorbance were determined at λ_max 620 nm using a plate reader (Bio-Tek Instruments ELx 800, Winooski, VT). Inhibition of recombinant Hsp90 was calculated as percent activity of the uninhibited control. DMSO concentrations were kept <1% in all experiments and controls. Samples and blanks were prepared in duplicates. Geldanamycin was tested as positive control, and negative controls were prepared by adding the substrate after reaction termination.

Results and discussion

Table 1 shows the inhibitory profiles of tested compounds against ATPase activity of Hsp90. Interestingly, torsemide, sulfadiazine, and sulfathiazole illustrated low micromolar IC₅₀ values, whereas sulfamethizole and sulfisoxazole exhibited mediocre IC₅₀ values. On the other hand, sulfadoxine, sulfaguanide, and sulfacetamide exhibited poor Hsp90 inhibition profiles.

Table 1.  Results of evaluated sulfonamides.

Drug	IC50 (µM) or % inhibition	pKa^b	Values of docking scoring functions^a
			Ligscore1^c	Ligscore2^c	−PLP1	−PLP2	Jain	−PMF	Consensus score
Torsemide	1.0 (0.94)^d	6.4	4.67	4.58	59.97	55.24	0.11	64.17	6^e
Sulfadiazine	1.5 (0.96)^d	6.5	4.17	3.89	55.53	49.24	2.62	52.8	6^e
Sulfathiazole	2.6 (0.99)^d	7.1	4.44	4.53	32.2	29.5	0.8	25.3	5^f
Sulfamethizole	50.1 (0.88)^g	5.5	4.21	4.9	36.1	31.31	1.13	31.0	6^f
Sulfisoxazole	60.7 (0.90)^g	5.0	3.67	4.62	46.5	40.13	−0.54	45.7	6^f
Sulfadoxine	25% at 100 µM	5.8	5.48	5.29	62.37	65.86	0.97	70.9	6^e
Sulfaguanide	11% at 100 µM	9.7	4.67	4.58	59.97	55.24	0.11	64.17	6^e
Sulfacetamide	9% at 100 µM	5.4	3.77	4.33	26	24.48	−0.38	32.9	4^f
Geldanamycin reference standard	272 nM (0.98)	—	—	—	—	—	—	—	—

^aSee references [45–51].

^bSee references [59,60].

^cCalculated using CFF force field.

^eFlexible docking.

^d,gValues in brackets represent the correlation coefficients of the corresponding dose–response lines determined over 3 or 4 concentration logarithmic cycles, respectively (i.e. 0.1, 1, 10, and 100 µM).

^fRigid docking.

To probe this intriguing behavior, we decided to dock the different sulfonamides into the ATP-binding pocket of Hsp90. The docking experiment was conducted by employing LIGANDFIT⁴⁵ and consensus scoring based on PLP1,⁴⁶ PLP2,⁴⁶ ligscore1,⁴⁷ ligscore2,⁴⁵ PMF,^48,49 and JAIN^50,51 scoring functions. However, since individual scoring function(s) are generally incapable of correctly evaluating binding free energies due to the high complexity of the underlying molecular interactions,^54–56 we decided to select optimal docked conformers/poses based on consensus among the six scoring functions.^57,58 The highest ranking docked conformer/pose was selected to represent the bound ligand. The consensus function assigned a value of 1 for any molecular pose ranked within the highest 40% by the particular scoring function; otherwise, it was assigned a zero value, that is, if it was within the lowest 60%. Subsequently, the consensus function summed up the scores for each molecular pose/conformer and ranked the molecular orientations accordingly. Consensus scoring should alleviate the inability of scoring functions to individually evaluate binding free energies correctly.^59–64

Table 1 shows the score values, including consensus scores, of optimal docked conformers/poses of each tested sulfonamide.

Figure 2 shows the docked poses of sulfonamides into the binding pocket of Hsp90. Figure 3 shows the binding interactions of torsemide and sulfadiazine as representatives of remaining sulfonamides.

Figure 2.  Docked poses of tested sulfonamides in Hsp90 (PDB code: 1YET, resolution 1.9 Å). (A) and (B) showing the binding pocket with and without solvent accessible surface, respectively.

Figure 3.  Docked poses of (A) torsemide and (B) sulfadiazine in Hsp90.

Apparently, all docked compounds, except sulfaguanide and sulfacetamide, share four critical features/interactions within the binding pocket: (i) an electron-deficient N-heterocycle stacking against the sulfide of MET98, (ii) a hydrogen-bonding interaction connecting the sulfonamidic NH (or aniline NH in torsemide) to the carboxylate of ASP54, (iii) a hydrogen-bonding interaction connecting the quaternary ammonium of LYS58 and sulfone oxygens in the ligands, and (iv) hydrogen-bonding interaction connecting aniline nitrogen (or urea NH in torsemide) to the carboxylate of ASP102.

The fact that potent and moderate inhibitors assume similar docking-based interactions suggested another factor to explain the apparent differences in activity. Upon evaluating the pK_a values of tested sulfonamides,^65,66 it became clear that potent inhibitors (torsemide, sulfadiazine, and sulfathiazole) are, on average, 1.2 pK_a units higher than moderate inhibitors (sulfamethizole, sulfisoxazole, and sulfadoxine),^65,66 suggesting certain role played by the ionizability of sulfonamidic NH in bioactivity.

Apparently, the greater acidity of moderate sulfonamides promotes ionization of sulfonamidic NH under pH 7.4 (bioassay pH) and therefore causes a loss of the critical hydrogen-bonding interaction with ASP54. Furthermore, ionization renders the sulfonamidic fragments negatively charged and hence electrostatically repulsive with ASP54 carboxylate.

Sulfadoxine seems to have another detrimental structural factor related to its inferior activity: its 4,5-dimethoxy-pyrimidine ring has higher electron density compared with corresponding heterocyclic rings in the other sulfonamides because of the electron-donating dimethoxy substituents. This property probably reduces the efficiency of π-stacking against the electron-rich sulfide of MET98 causing further reduction in bioactivity, that is, compared with sulfamethizole and sulfisoxazole.

The poor inhibitory action of sulfacetamide is also explainable by the high acidity of its sulfonamidic NH due to its mixed amidic/sulfonamidic nature. Furthermore, sulfacetamide lacks electronically deficient heterocycle capable of stacking against MET98 sulfide, as shown in Figure 4A.

Figure 4.  Docked poses of (A) sulfacetamide and (B) sulfaguanide in Hsp90.

Finally, sulfaguanide seems to assume flipped pose (Figure 4B): its aniline NH₂ is suboptimally hydrogen-bonded to ASP93 (at a distance of 4.25 Å), whereas the sulfonamidic oxygens are hydrogen-bonded to LYS58. Similarly, the terminal guanidino seems to interact with the carboxylate of ASP102. However, although the aromatic ring in sulfaguanide stacks against MET98 sulfur, the fact that it is electron-rich aniline suggests weaker attraction with MET98 sulfide compared with electron-deficient heterocycles in sulfadiazine, sulfathiazole, and torsemide. Furthermore, this flipped pose cost sulfaguanide a critical hydrogen-bonding interaction with ASP54. The critical significance of this interaction can be deduced from the consensus of other more potent sulfonamides (e.g. torsemide, sulfadiazine, and sulfathiazole).

Moreover, the alkaline nature of the guanidine moiety of sulfaguanide (pK_a = 9.7)^65,66 suggests that it is positively charged under bioassay conditions (pH = 7.4) and hence should be heavily hydrated, particularly as it resides at the outer rim of the binding site. This factor probably further contributes to the poor Hsp90 inhibition of sulfaguanide.

These structure–activity trends suggest that other sulfonamide-related linkers can be useful to generate and optimize new potent Hsp90 inhibitors, for example, sulfonylurea. Furthermore, it is anticipated that introduction of small amidic substituents at the para-position of the aromatic heterocycle (e.g. pyrimidine in sulfadiazine) should allow hydrogen-bonding interactions with ASP93 carboxylate without increasing the electron density of the heterocycle, which might jeopardize aromatic stacking against MET98 sulfide.

Conclusion

The current research shows promising Hsp90 inhibition for some clinically established sulfonamides. Furthermore, docking experiments illustrated the significance of four critical amino acids in binding to Hsp90, namely MET98, ASP54, ASP102, and LYS58. Our findings and explanations pave the way for subsequent optimization of new potent sulfonamidic Hsp90 inhibitors.

Declaration of interest

The authors thank the Deanship of Scientific Research and Hamdi-Mango Center for Scientific Research at the University of Jordan for their generous funds. The authors also thank Al-Zaytoonah Private University of Jordan for support.