Development of Covalent Ligand–Receptor Pairs to Study the Binding

Letters pubs.acs.org/acschemicalbiology

Development of Covalent Ligand−Receptor Pairs to Study the Binding Properties of Nonpeptidic Neurotensin Receptor 1 Antagonists Ralf C. Kling, Manuel Plomer, Christopher Lang, Ashutosh Banerjee, Harald Hübner, and Peter Gmeiner* Department of Chemistry and Pharmacy, Emil Fischer Center, Friedrich Alexander University, Schuhstr. 19, 91052 Erlangen, Germany S Supporting Information *

ABSTRACT: The neurotensin receptor NTS1 has been suggested to be of pharmaceutical relevance, as it was found to exert modulatory effects on dopaminergic signal transduction and to be involved in tumor progression. Rational drug design of NTS1 receptor ligands requires molecular insights into the binding behavior of a particular lead compound. Although crystal structures of NTS1 have revealed the molecular determinants of peptide−agonist interactions, the binding mode of small-molecule antagonists remains largely unknown. Employing a disulfide-based tethering approach, we developed covalently binding molecular probes. The ligands 1 and 2 are based on the pharmacophore of the nonpeptidic NTS1 antagonist SR142948A, allowing the formation of a disulfide bond to an engineered cysteine residue of NTS1. The position of the covalent bond between Cys1272.65 and the ligand was used to predict the binding mode of the covalent antagonist 1 and its parent compound SR142948A by molecular dynamics simulations.

G

conformations will facilitate our understanding of the molecular events that lead to receptor activation. However, crystallization of the NTS1 receptor bound to an antagonist has not been described, yet. It has been shown that computational methods including molecular docking and molecular-dynamics (MD) simulations can be applied to determine binding modes and ligand− receptor interactions.14−17 However, the prediction of reliable binding poses for molecules that exhibit substantial structural differences compared to cocrystallized ligands is highly challenging. The use of tailor-made ligands capable of forming a covalent bond with suitable residues of the receptor will be very helpful to determine the overall position and orientation of an individual ligand chemotype and to subsequently predict reliable binding modes by computational methods. When a disulfide-based tethering approach is employed, which has been shown previously to be applicable to a variety of different proteins including GPCRs,18,19 the specificity of cross-linking depends on substantial affinity of the pharmacophore to the target receptor. If the position of the reactive moiety of the ligand is close to a nucleophilic cysteine residue, specific formation of a covalent bond by disulfide transfer is possible.20 We have shown that disulfide-based covalent ligands represent excellent tools for structural studies on GPCRs, which facilitate crystallization.21,22 Using the β2-adrenoceptor H932.64C mutant (β2AR-H932.64C, superscripts refer to the generic Ballesteros−

-protein-coupled receptors (GPCRs) are involved in a wide range of physiological functions and are thus considered as highly attractive targets for drug design.1 The last couple of years have seen remarkable progress in structural biology,2,3 which led the way to the unveiling of a variety of different GPCR crystal structures.4 As part of this success, the molecular architecture of an engineered construct of rat neurotensin receptor NTS1 coupled to NT(8−13) (H-ArgArg-Pro-Tyr-Ile-Leu-OH), the C-terminal hexapeptide of the endogenous tridecapeptide neurotensin, was elucidated.5 Subsequently, additional crystal structures of this receptor have been published corroborating the interactions between NT(8−13) and NTS1.6,7 The NTS1 receptor has been suggested to be of pharmaceutical relevance, as it was found to be coexpressed with dopamine receptors on neurons8 and thus has been described to exert modulatory effects on dopaminergic signal transduction.9,10 Various efforts in medicinal chemistry led to the development of the nonpeptide antagonist SR48692, an NTS1 selective drug candidate,11 and a [18F]-labeled diarylpyrazole glycoconjugate representing a highly promising tracer for the positron-emission-tomography (PET) of tumors overexpressing the NTS1 receptor.12 Although the present crystal structures of NTS1 have provided valuable insights into the molecular determinants of peptide− agonist interactions, the binding mode of small-molecule antagonists remains largely unknown. Detailed knowledge about NTS1−antagonist interactions will guide the design of subtype selective antagonists, nonpeptidic agonists, as well as tracers for PET and radioactive tumor therapeutics.13 Moreover, the comparison of agonist- and antagonist-bound NTS1 © XXXX American Chemical Society

Received: November 20, 2015 Accepted: January 25, 2016

A

DOI: 10.1021/acschembio.5b00965 ACS Chem. Biol. XXXX, XXX, XXX−XXX

Letters

ACS Chemical Biology

Figure 1. Overall strategy to investigate nonpeptidic antagonists at NTS1 including the generation of a covalent ligand−receptor pair and a cysteine mutant of NTS1 as well as molecular dynamics (MD) simulations.

Scheme 1. Synthesis of Compounds 1 and 2a

a Reagents and conditions: (a) amine, Mo(CO)6, Herrmann’s palladacycle, [(t-Bu)3PH]BF4, DBU, 1,4-dioxane, 140 °C, microwave, 30 min, 71% (ref 24). (b) (1) 1,4-dithiane-2,5-diol, NaBH(OAc)3, THF, rt, 24 h; (2) 2 N HCl, 2-mercaptoethanol, 1,3-dibromo-5,5-dimethylhydantoin, DCM, rt, 1 h, 34%. (c) (1) Mo(CO)6, Herrmann’s palladacycle, [(t-Bu)3PH]BF4, amine, DBU, 1,4-dioxane, 140 °C, 30 min microwave, 50%; (2) TBAF, THF, rt, 3 h, 83% (ref 12). (d) CuSO4·5H2O, sodium ascorbate, t-BuOH, H2O, rt, 48 h, 11%. (e) (1) H2SO4 conc., HBr (48 wt % in H2O), 0 °C to rt, 24 h, 100 °C, 3 h, 85%; (2) NaN3, DMF, 80 °C, 7 h, 60%.

Weinstein numbering system for GPCRs) as a prototype, we found that the bioactive conformation of the pharmacophores and the overall structure of the receptor were very similar when compared to complexes between the diffusible ligand analogs and wild-type receptor.22 The extracellular end of transmembrane helix (TM) 2 proved to be an excellent region for the formation of a covalent bond with a disulfide-functionalized ligand. Beyond adrenoceptors, the homologous position in TM2 could also be used for the construction of functional

GPCR−ligand pairs between dopamine, serotonin, and histamine receptors and their covalent neurotransmitter analogs.22 Extending our strategy toward peptidic GPCRs, we herein describe development of covalently binding molecular probes for the neurotensin receptor subtype 1 (NTS1), which are based on the pharmacophore of the nonpeptidic antagonist SR142948A. Cross-linking was facilitated by the attachment of a linker with a reactive thiol to the scaffold of SR142948A, B

DOI: 10.1021/acschembio.5b00965 ACS Chem. Biol. XXXX, XXX, XXX−XXX

Letters

ACS Chemical Biology

the radioligand [3H]NT(8−13) (Table S1). For the mutants Y1.39C and F2.65C, we detected only a minor impact on expression levels and binding characteristics, as both Bmax and KD values for [3H]NT(8−13) showed only negligible differences compared to wild type (Table S1). In contrast, we observed a 10-fold reduced KD for E2.61C-NTS1 and a negative influence on the expression level (20-fold reduction). To investigate binding affinity of the test compounds 1 and 2 and their lead compound SR142948A, we performed radioligand-displacement assays (Table S1). Using a wild type receptor, both test compounds exhibited Ki values in the single digit nanomolar range (1, Ki = 4.0 nM; 2, Ki = 5.5 nM), whereas SR142948A gave a Ki of 0.59 nM. Interestingly, the Ki values of 2.7 nM and 6.5 nM for the disulfides 1 and 2, respectively, at the NTS1 mutant F2.65C indicated that binding affinity was not significantly impaired by the amino acid exchange. Encouraging data were also obtained for the backup receptor mutant E2.61C when the Ki values conferred an approximately 3-fold lower affinity for the test compounds and the reference agent than for F2.65C. The Y1.39C mutation gave reduced affinity of all three ligands (7- to 19-fold). The capacity of the single point mutants to be activated by the endogenous agonist neurotensin was investigated using a luciferase reporter-gene assay. Our results show that neurotensin was able to activate wild type NTS1 and the three receptor mutants investigated (Figures S2 and S3). Employing high concentrations of neurotensin (1 μM), the ratio of the agonist mediated effect and basal activity of NTS1-F2.65C was highly similar to wild type NTS1 (2.6 fold over basal and 2.7 fold over basal, respectively, Figure S2). Although we observed an increased basal activity for the backup candidates E2.61C and Y1.39C, they still showed significant ligand-specific activation in the presence of neurotensin (Figure S2). Dose− response curves leading to EC50 values of 1.8 nM and 1.3 nM confirmed the high similarity between NTS1-F2.65C and the wild type receptor, respectively, toward agonist stimulation (Figure S3, Table S2). The alternative mutants E2.61C and Y1.39C showed a significantly and dose-dependently reduced potency (EC50 = 18 nM and 6.3 nM, respectively). To verify that our test compounds 1 and 2 behave as antagonists, we tested their ability to inhibit the activation of wild type and mutant NTS1 receptors in the presence of 30 nM neurotensin using the MAPK-driven functional assay (Figure S4). As expected, the covalent candidates 1 and 2 as well as the diffusible parent compound SR142948A27 were capable of blocking the reporter gene signal initiated by neurotensin in a dose-dependent manner. Whereas antagonist properties were observed for F2.65C and the backup candidate Y1.39C, 1, 2, and SR142948A revealed inverse agonist effects for NTS1E2.61C significantly attenuating basal activity (Figure S4), the latter of which being compatible with its increased basal activity (Figure S2). On the basis of a competitive radioligand depletion assay, we investigated the capacity of the ligand−receptor complex to form a covalent bond following a recently established protocol.22 F2.65C-NTS1 was considered the most promising candidate, because its expression level and activation properties were most similar to wild type NTS1 and thus was chosen for these experiments. In detail, membranes from HEK293 cells containing either the F2.65C-NTS1 mutant or the wild type receptor were incubated with the desired test compound. In addition to the putatively covalent compounds 1 and 2, containing a short and a long linker between the identical

which allows the formation of a disulfide bond to cysteine mutants of the NTS1 receptor. This covalent bond should provide a structural starting point for subsequent computational studies designed to investigate the binding mode of SR142948A-derived antagonists at NTS1. Employing our recent methodology, we envisioned the construction of a covalent ligand−receptor pair from a disulfide-functionalized ligand and a cysteine mutant of NTS1 (Figure 1). The disulfide-based strategy was preferred over other cross-linking reactions including Michael additions or nitrogen-mustard-promoted alkylations, because of the chemoselectivity of the reactive thiol moiety for a suitably positioned cysteine residue over other nucleophilic amino acids side chains.20 As the crystal structures of NTS1 indicated a lack of targetable cysteine residues in proximity of the binding pocket, we envisioned the introduction of point mutations at appropriate positions of the receptor. Suitable candidates should exhibit residue side chains pointing into the binding pocket and be located within the helical domains to reduce their conformational flexibility. Encouraged by the ability of the X2.64C mutants of the monoaminergic GPCRs β2AR, D2R, 5HT2A, and H1,22 TM2 of NTS1 was chosen for single point mutations. The homologous residue Asn1262.64 in NTS1 was not mutated to cysteine, as the side chain was not pointing into the binding pocket according to the crystal structure (Figure S1). Instead, we focused on mutating the adjacent solventexposed residue Phe1272.65, for which we previously found that its substitution with alanine impaired neither the expression levels of NTS1 nor the binding of antagonists.23 To access two suitable backup candidates, we exchanged the proximate residues Glu1232.61 and Tyr701.39 for cysteine (Figure S1). In the following, these mutants are referred to as Y1.39C, E2.61C, and F2.65C. To be able to address these mutants, the nonpeptidic test compounds 1 and 2 were synthesized, which share a common SR142948A-derived core pharmacophore but differ in the linker length. Thus, disulfide 1 has a short ethylene linker between the tertiary amine and the disulfide function, whereas the click chemistry ligation product 2 incorporates a nine-atom spacer (Scheme 1). The linker units were positioned at the amine moiety of the lead compound, as previous binding data suggested only a minor influence of structural modifications on NTS1 binding.12 The target compounds 1 and 2 were synthesized starting from the heterocyclic building block 3 (Scheme 1), which was prepared according to our recently described protocol.24 Thus, a palladium-catalyzed aminocarbonylation reaction using N,N′dimethyl propylenediamine yielded the secondary amine 4. Subsequently, reductive alkylation was done using the aldehyde-synthon 1,4-dithiane-2,5-diol and sodium triacetoxyborohydride. After hydrolysis of the crude product, an excess of 2-mercaptoethanol and the oxidizing agent 1,3-dibromo-5,5dimethylhydanthoin25 were added to avoid homodimerization and direct the reaction to a preferred formation of the asymmetric disulfide 1. The triazole-based target compound 2 was obtained from the N-hexynyl derivative 5, which was synthesized by aminocarbonylation of the aryl bromide 3.12 Copper-catalyzed alkyne−azide cycloaddition (CuAAC) with bis(2-azidoethyl)disulfide (6)26 resulted in formation of the disulfide 2. Expression levels and binding properties of the single point mutants Y1.39C, E2.61C, and F2.65C were determined and compared to wild type NTS1. Thus, HEK293 cells were transiently transfected and tested for saturation binding using C

DOI: 10.1021/acschembio.5b00965 ACS Chem. Biol. XXXX, XXX, XXX−XXX

Letters

ACS Chemical Biology

Figure 2. Radioligand depletion assay to investigate covalent ligand binding at the NTS1 mutant F2.65C. Labeling of the free binding sites was accomplished using (A) [3H]NT(8−13) and (B) [3H]SR48692. Measured radioactivity has been normalized on unspecific and total binding being 0% and 100%, respectively. Data represent average values + SEM obtained from five independent experiments, each performed as triplicates.

Figure 3. Ligand−receptor interactions for the covalently tethered compound 1 (blue, A and B) and SR142948A (cyan, C and D) at NTS1-F2.65C and NTS1 wild type, respectively. Both mutant and wild type NTS1 are shown in gray (A, C). Residues stabilizing the conformation of 1 and SR142948A are labeled. The Ballesteros−Weinstein numbering has been used for transmembrane (TM) residues, whereas sequence numbers are given for residues of extracellular loops (EL) and the N-terminus (N-term.). Individual figures have been prepared using UCSF Chimera (A, C) and the modified output of LigPlot+ (B, D).31,32

pharmacophore and the reactive disulfide moiety, respectively, we used the reversible parent compound SR142948A as a

negative control. Neither with NTS1 wild type nor with SR142948A should covalent bond formation be observed. All D

DOI: 10.1021/acschembio.5b00965 ACS Chem. Biol. XXXX, XXX, XXX−XXX

Letters

ACS Chemical Biology

which participate in stabilizing NT(8−13) by hydrogen bond interactions (Figure S6). Moreover, we observed a conformational change of EL3, including extracellular parts of TM7 and, to a minor extent, TM6, which in turn influenced the conformation of Trp334 of EL3 (Figure S6). Together with other residues from the upper part of TM7, Trp334 of EL3 was shown to be crucial for the signal propagation of rat NTS1 and the coupling to G proteins.30 The disulfide-bearing linker domain is surrounded by residues of the N-terminus, EL1 and EL2, and bends back into the binding pocket, thus enabling the formation of the covalent disulfide bond with Cys1272.65 of TM2. We believe that being localized within this conformationally flexible subpocket between the N-terminus, EL1, and EL2 is likely to also allow the covalent linkage of 2 by accommodating to its longer linker. In general, the conformation of 1 described above is compatible with available mutational data on the binding mode of the closely related antagonist SR48692 at rat NTS1, attributing a key role in ligand recognition to residues Met208 of EL2, Tyr3246.51, Arg3276.54, Phe3316.58, and Tyr3517.35 (which correspond to Met207EL2, Tyr3196.51, Arg3226.54, Phe3266.58, and Tyr3467.35 in human NTS1).28 Among these residues, Met207EL2, Arg3276.54, Phe3266.58, and Tyr3467.35 were found to form direct interactions with 1 (Figure 3), whereas Tyr3196.51 is positioned directly below Arg1483.32 in TM3 and Arg3226.54/Arg3236.55 in TM6 and is therefore likely to influence binding of 1 by modulating the conformation of those amino acids. To extend our studies toward a diffusible ligand, we replaced the covalent molecular probe 1 with the diffusible parent compound SR142948A in each simulation system, restored the wild type sequence of NTS1 by mutating Cys1272.65 to phenylalanine, and submitted the three SR142948A−NTS1 complexes to additional MD simulation runs of 500 ns each. The global conformation of SR142948A within each run remained stable throughout the simulated time scale (Figure S5B) and was virtually identical to that of the tethered antagonist (Figure 3C,D). However, we observed an increased degree of conformational flexibility of the terminal alkylamine substituent of SR142948A, suggesting only a minor contribution of the latter ligand domain to the overall affinity of the ligand. In one of the simulation runs, the dimethylamine domain even started to gain access to extracellular receptor domains (Figure S5B), which is in agreement with previous experimental studies that showed only a small influence of structural modifications at that position on NTS1 binding.12 As for the 1-bound systems, we were able to detect similar conformational differences at EL3 and residues on the extracellular half of TM7 (Figure S6). In conclusion, we extended our covalent ligand approach21,22 toward the neurotensin receptor NTS1 in order to determine the overall position and orientation of SR142948A-type ligands. Employing a disulfide-based tethering, we developed the covalently binding molecular probes 1 and 2 for NTS1, which are based on the pharmacophore of the nonpeptidic antagonist SR142948A. Mutation of a phenylalanine in position 1272.65 of human NTS1 by cysteine led to an engineered receptor with almost wild-type properties. The attachment of a linker with a reactive thiol to the scaffold of SR142948A enabled the generation of disulfide-based, covalent ligand− receptor pairs. The high binding affinity of the parent compound with Ki values in the nanomolar range was helpful in generating a specific candidate. Nevertheless, the method allows also the design of specific covalent ligands based on

compounds were applied at a concentration of 10 nM. After centrifugation (to stop incubation) and washing, the radioligands [3H]NT(8−13) and [3H]SR48692 were added in parallel experiments. The results of the radioligand depletion assay with NTS1 wild type showed that incubation with SR142948A, 1, and 2 did not yield an irreversible occupation of the binding pocket, as both radioligands, [3H]NT(8−13) and [3H]SR48692, were able to displace those compounds (Figure 2). At the F2.65C mutant, the majority of originally available binding sites was occupied by the covalent test compounds 1 and 2 due to the formation of a covalent ligand−receptor complex, thus preventing the binding of the radioligands, whereas it was still possible to radiolabel the binding sites of the F2.65C mutant upon preincubation with the diffusible compound SR142948A. Using [3H]NT(8−13) as a radioligand, 39% and 25% of the originally available receptor binding sites could be specifically addressed after incubation with compounds 1 and 2, respectively, whereas we observed values of 21% and 19% for 1 and 2, respectively, employing [3H]SR48692 as a radioligand. Therefore, our results demonstrate that both 1 and 2 are able to form covalent ligand−receptor complexes with F2.65C-NTS1. To learn more about the binding mode of SR142948Aderived antagonists at NTS1, we took advantage of our covalent molecular probe 1 and performed MD simulations of the disulfide-bridged ligand−receptor complex. Moreover, we used the results of these simulations to perform additional simulation runs of SR142948A coupled to the NTS1 wild type. As it was not possible to obtain an initial conformation of 1 by conventional molecular docking, 1 was placed into the NTS1F2.65C mutant by hand, thereby taking advantage of the knowledge of the covalent interaction between 1 and the mutated cysteine residue. In this initial pose, the carboxylate moiety of 1 formed an ionic interaction with Arg3226.54 of TM6 (a residue, which was shown to be essential for the binding of SR-type antagonists28), when the reactive disulfide group was in close proximity to residue Cys1272.65 of TM2. After having tethered 1 to F2.65C (a more detailed description of this procedure is provided in the Methods section), three independent MD simulation runs of 500 ns each were carried out to examine the conformational stability of this covalent ligand−receptor complex. RMSD analyses indicated high conformational stability of 1 over the investigated time period (Figure S5A). Moreover, we have observed a virtually identical binding pose of 1 throughout the individual simulation runs (Figure S5A). In those poses, the adamant moiety of 1 occupies a cavity comprised of residues from TM3, TM5 to TM7, and EL2, when its carboxylate is stabilized by hydrogen-bond/ionic interactions to Arg1483.32, Arg3226.54, and Arg3236.55 (Figure 3A,B). It is important to note that the aforementioned interactions represent the only stable hydrogen bonds between 1 and the receptor, which is in contrast to the larger number of hydrogen bonds between NT(8−13) and NTS1 observed within both the crystal structures of NTS15−7 and our previous simulations of NT(8−13) coupled to NTS1.29 The central pyrazole fragment positions the dimethoxyphenyl substituent in proximity to TM6, TM7, and EL3 and the isopropyl-phenyl residue into the direction of the extracellular entrance of the binding pocket (Figure 3A,B). The sterically demanding dimethoxyphenyl moiety adopts a position similar to Pro10 of NT(8−13). Compared to the crystal structure of NT(8− 13)-coupled NTS1, 1 leads to a displacement of residues Tyr3427.31 and His3437.32 on the upper part of TM7, both of E

DOI: 10.1021/acschembio.5b00965 ACS Chem. Biol. XXXX, XXX, XXX−XXX

Letters

ACS Chemical Biology

(3) Ghosh, E., Kumari, P., Jaiman, D., and Shukla, A. K. (2015) Methodological advances: the unsung heroes of the GPCR structural revolution. Nat. Rev. Mol. Cell Biol. 16, 69−81. (4) Shonberg, J., Kling, R. C., Gmeiner, P., and Löber, S. (2015) GPCR crystal structures: Medicinal chemistry in the pocket. Bioorg. Med. Chem. 23, 3880−3906. (5) White, J. F., Noinaj, N., Shibata, Y., Love, J., Kloss, B., Xu, F., Gvozdenovic-Jeremic, J., Shah, P., Shiloach, J., Tate, C. G., and Grisshammer, R. (2012) Structure of the agonist-bound neurotensin receptor. Nature 490, 508−513. (6) Egloff, P., Hillenbrand, M., Klenk, C., Batyuk, A., Heine, P., Balada, S., Schlinkmann, K. M., Scott, D. J., Schutz, M., and Pluckthun, A. (2014) Structure of signaling-competent neurotensin receptor 1 obtained by directed evolution in Escherichia coli. Proc. Natl. Acad. Sci. U. S. A. 111, 655−662. (7) Krumm, B. E., White, J. F., Shah, P., and Grisshammer, R. (2015) Structural prerequisites for G-protein activation by the neurotensin receptor. Nat. Commun. 6, 7895. (8) Kinkead, B., Binder, E. B., and Nemeroff, C. B. (1999) Does neurotensin mediate the effects of antipsychotic drugs? Biol. Psychiatry 46, 340−351. (9) Binder, E. B., Kinkead, B., Owens, M. J., and Nemeroff, C. B. (2001) Neurotensin and dopamine interactions. Pharmacol Rev. 53, 453−486. (10) Koschatzky, S., Tschammer, N., and Gmeiner, P. (2011) Crossreceptor interactions between dopamine D2L and neurotensin NTS1 receptors modulate binding affinities of dopaminergics. ACS Chem. Neurosci. 2, 308−316. (11) Gully, D., Canton, M., Boigegrain, R., Jeanjean, F., Molimard, J. C., Poncelet, M., Gueudet, C., Heaulme, M., Leyris, R., Brouard, A., et al. (1993) Biochemical and pharmacological profile of a potent and selective nonpeptide antagonist of the neurotensin receptor. Proc. Natl. Acad. Sci. U. S. A. 90, 65−69. (12) Lang, C., Maschauer, S., Hubner, H., Gmeiner, P., and Prante, O. (2013) Synthesis and evaluation of a (18)F-labeled diarylpyrazole glycoconjugate for the imaging of NTS1-positive tumors. J. Med. Chem. 56, 9361−9365. (13) Maschauer, S., Ruckdeschel, T., Tripal, P., Haubner, R., Einsiedel, J., Hubner, H., Gmeiner, P., Kuwert, T., and Prante, O. (2014) In vivo monitoring of the antiangiogenic effect of neurotensin receptor-mediated radiotherapy by small-animal positron emission tomography: a pilot study. Pharmaceuticals 7, 464−481. (14) Leach, A. R., Shoichet, B. K., and Peishoff, C. E. (2006) Prediction of protein-ligand interactions. Docking and scoring: successes and gaps. J. Med. Chem. 49, 5851−5855. (15) Shoichet, B. K., and Kobilka, B. K. (2012) Structure-based drug screening for G-protein-coupled receptors. Trends Pharmacol. Sci. 33, 268−272. (16) Dror, R. O., Pan, A. C., Arlow, D. H., Borhani, D. W., Maragakis, P., Shan, Y., Xu, H., and Shaw, D. E. (2011) Pathway and mechanism of drug binding to G-protein-coupled receptors. Proc. Natl. Acad. Sci. U. S. A. 108, 13118−13123. (17) Dror, R. O., Green, H. F., Valant, C., Borhani, D. W., Valcourt, J. R., Pan, A. C., Arlow, D. H., Canals, M., Lane, J. R., Rahmani, R., Baell, J. B., Sexton, P. M., Christopoulos, A., and Shaw, D. E. (2013) Structural basis for modulation of a G-protein-coupled receptor by allosteric drugs. Nature 503, 295−299. (18) Erlanson, D. A., Wells, J. A., and Braisted, A. C. (2004) Tethering: fragment-based drug discovery. Annu. Rev. Biophys. Biomol. Struct. 33, 199−223. (19) Buck, E., and Wells, J. A. (2005) Disulfide trapping to localize small-molecule agonists and antagonists for a G protein-coupled receptor. Proc. Natl. Acad. Sci. U. S. A. 102, 2719−2724. (20) Weichert, D., and Gmeiner, P. (2015) Covalent Molecular Probes for Class A G Protein-Coupled Receptors: Advances and Applications. ACS Chem. Biol. 10, 1376−1386. (21) Rosenbaum, D. M., Zhang, C., Lyons, J. A., Holl, R., Aragao, D., Arlow, D. H., Rasmussen, S. G., Choi, H. J., Devree, B. T., Sunahara, R. K., Chae, P. S., Gellman, S. H., Dror, R. O., Shaw, D. E., Weis, W. I.,

pharmacophores with micromolar Ki values as we have shown for disulfide-functionalized analogs of biogenic amines.22 For such examples, a careful design of the linker providing additional binding energy is helpful. The knowledge about this particular covalent bond was used to facilitate the prediction of reliable binding modes for the covalent ligand 1 and SR142948A by long-term MD simulations. Molecular interactions between the neurotensin receptors and nonpeptidic ligands have been investigated.28,33 This work presents the first model of an antagonist-bound human NTS1 that combines an experimental with a computational approach. Taken together, our strategy has provided us with a plausible structural basis for the binding mode of SR142948A-derived antagonists at NTS1, which may facilitate a rational, structurebased development of subtype selective antagonists, nonpeptidic agonists, as well as tracers for PET and radioactive tumor therapeutics.

■

METHODS

■

ASSOCIATED CONTENT

Full details for all experimental methods (site-directed mutagenesis, saturation-binding experiments and receptor binding studies, radioligand depletion assay, luciferase reporter-gene assay, synthesis protocols, and computational methods) are provided in the Supporting Information.

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acschembio.5b00965. Figures S1−S6, Tables S1 and S2, complete experimental procedures for the biological characterization of the test compounds and NTS1 receptor mutants (site-directed mutagenesis, saturation-binding experiments and receptor binding studies, radioligand depletion assay, and luciferase reporter-gene assay), synthesis protocols, and a detailed description of computational methods (PDF) Coordinates of representative, energy-minimized snapshots of compound 1 and SR142948A in complex with NTS1-F2.65C (PDB) Coordinates of representative, energy-minimized snapshots of compound 1 and SR142948A in complex with NTS1 wild type (PDB)

■

AUTHOR INFORMATION

Corresponding Author

*Tel.: +49 9131 85-29383. Fax: +49 9131 85-22585. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank the German Research Foundation (DFG grant Gm13/10).

■

REFERENCES

(1) Wise, A., Gearing, K., and Rees, S. (2002) Target validation of Gprotein coupled receptors. Drug Discovery Today 7, 235−246. (2) Kobilka, B. (2013) The structural basis of g-protein-coupled receptor signaling (nobel lecture). Angew. Chem., Int. Ed. 52, 6380− 6388. F

DOI: 10.1021/acschembio.5b00965 ACS Chem. Biol. XXXX, XXX, XXX−XXX

Letters

ACS Chemical Biology Caffrey, M., Gmeiner, P., and Kobilka, B. K. (2011) Structure and function of an irreversible agonist-beta(2) adrenoceptor complex. Nature 469, 236−240. (22) Weichert, D., Kruse, A. C., Manglik, A., Hiller, C., Zhang, C., Hubner, H., Kobilka, B. K., and Gmeiner, P. (2014) Covalent agonists for studying G protein-coupled receptor activation. Proc. Natl. Acad. Sci. U. S. A. 111, 10744−10748. (23) Harterich, S., Koschatzky, S., Einsiedel, J., and Gmeiner, P. (2008) Novel insights into GPCR-peptide interactions: mutations in extracellular loop 1, ligand backbone methylations and molecular modeling of neurotensin receptor 1. Bioorg. Med. Chem. 16, 9359− 9368. (24) Lang, C., and Gmeiner, P. (2013) Efficient Synthesis of Heterocyclic Neurotensin Receptor Ligands by Microwave-Assisted Aminocarbonylation. Synthesis 45, 2474−2480. (25) Khazaei, A., Zolfigol, M. A., and Rostami, A. (2004) 1,3dibromo-5,5-dimethylhydantoin [DBDMH] as an efficient and selective agent for the oxidation of thiols to disulfides in solution or under solvent-free conditions. Synthesis, 2959−2961. (26) Zhang, S. Y., and Zhao, Y. (2010) Rapid Release of Entrapped Contents from Multi-Functionalizable, Surface Cross-Linked Micelles upon Different Stimulation. J. Am. Chem. Soc. 132, 10642−10644. (27) Einsiedel, J., Held, C., Hervet, M., Plomer, M., Tschammer, N., Hubner, H., and Gmeiner, P. (2011) Discovery of highly potent and neurotensin receptor 2 selective neurotensin mimetics. J. Med. Chem. 54, 2915−2923. (28) Labbe-Jullie, C., Barroso, S., Nicolas-Eteve, D., Reversat, J. L., Botto, J. M., Mazella, J., Bernassau, J. M., and Kitabgi, P. (1998) Mutagenesis and modeling of the neurotensin receptor NTR1. Identification of residues that are critical for binding SR 48692, a nonpeptide neurotensin antagonist. J. Biol. Chem. 273, 16351−16357. (29) Schaab, C., Kling, R. C., Einsiedel, J., Hübner, H., Clark, T., Seebach, D., and Gmeiner, P. (2014) Structure-Based Evolution of Subtype-Selective Neurotensin Receptor Ligands. ChemistryOpen 3, 206−218. (30) Richard, F., Barroso, S., Nicolas-Etheve, D., Kitabgi, P., and Labbe-Jullie, C. (2001) Impaired G protein coupling of the neurotensin receptor 1 by mutations in extracellular loop 3. Eur. J. Pharmacol. 433, 63−71. (31) Pettersen, E. F., Goddard, T. D., Huang, C. C., Couch, G. S., Greenblatt, D. M., Meng, E. C., and Ferrin, T. E. (2004) UCSF Chimera - a visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605−1612. (32) Laskowski, R. A., and Swindells, M. B. (2011) LigPlot+: multiple ligand-protein interaction diagrams for drug discovery. J. Chem. Inf. Model. 51, 2778−2786. (33) Lückmann, M., Holst, B., Schwartz, T. W., and Frimurer, T. M. (2016) In Silico Investigation of the Neurotensin Receptor 1 Binding Site: Overlapping Binding Modes for Small Molecule Antagonists and the Endogenous Peptide Agonist. Mol. Inf. 35, 19.

G

DOI: 10.1021/acschembio.5b00965 ACS Chem. Biol. XXXX, XXX, XXX−XXX