pharmaceuticals

Article New Insights into the Stereochemical Requirements of the BB1 Receptor Antagonists Binding

Bahareh Rasaeifar, Patricia Gomez-Gutierrez and Juan J. Perez * Department of Chemical Engineering, Universitat Politecnica de Catalunya, ETSEIB, Av. Diagonal, 647, 08028 Barcelona, Spain; [email protected] (B.R.); [email protected] (P.G.-G.) * Correspondence: [email protected]

 Received: 4 August 2020; Accepted: 12 August 2020; Published: 17 August 2020 

Abstract: Members of the family of bombesinlike exert a wide range of biological activities both at the central nervous system and in peripheral tissues through at least three G-Protein Coupled Receptors: BB1, BB2 and BB3. Despite the number of ligands already described, only a few small molecule binders have been disclosed so far, hampering a deeper understanding of their pharmacology. In order to have a deeper understanding of the stereochemical features characterizing binding to the BB1 receptor, we performed the molecular modeling study consisting of the construction of a 3D model of the receptor by homology modeling followed by a docking study of the peptoids PD168368 and PD176252 onto it. Analysis of the complexes permitted us to propose prospective bound conformations of the compounds, consistent with the experimental information available. Subsequently, we defined a pharmacophore describing minimal stereochemical requirements for binding to the BB1 receptor that was used in silico screening. This exercise yielded a set of small molecules that were purchased and tested, showing affinity to the BB1 but not to the BB2 receptor. These molecules exhibit scaffolds of diverse chemical families that can be used as a starting point for the development of novel BB1 antagonists.

Keywords: bombesin receptors; antagonism; G-protein coupled receptors homology modeling; non-peptide neuromedin B antagonists

1. Introduction Members of the bombesinlike family of peptides, originally isolated from the skin of diverse amphibians and later found to be widely distributed in mammals [1,2], are compounds with a wide spectrum of biological activity. Thus, in the central nervous system they are involved in satiety, control of circadian rhythm and thermoregulation, whereas in peripheral tissues, stimulation of release, macrophages activation and effects on development [3]. In addition, they are known to play a role in the control of cellular proliferation [4,5]. Actions of this family of peptides are mediated through at least three G-Protein Coupled Receptors: the neuromendin B receptor (known as BB1R), the -releasing peptide receptor (known as BB2R), and the orphan, since its endogenous has not been disclosed yet, subtype 3 (known as BB3R) [6–8]. Due to the wide spectrum of biological activities mediated by these receptors, there is considerable interest in understanding their potential use as therapeutic agents. A literature review reveals potential therapeutic use for both agonists and antagonists targeting any of the three receptors for cancer therapy [9,10]: BB3R agonists for the treatment of obesity/diabetes mellitus [11]; BB2R antagonists for the treatment of radiation-induced lung injury [12] and BB1R or BB2R antagonists for the treatment of itching in atopic dermatitis [13]. Neuromedin B (NMB) and the gastrin-releasing peptide (GRP), together with its shorter version GRP(18–27)—known as neuromedin C (NMC)—are the mammal endogenous ligands of the BB1R and

Pharmaceuticals 2020, 13, 197; doi:10.3390/ph13080197 www.mdpi.com/journal/pharmaceuticals Pharmaceuticals 2020, 13, x FOR PEER REVIEW 2 of 16

Pharmaceuticals 2020, 13, 197 2 of 16 and BB2R, respectively [1,2]. They are selective agonists for the respective receptors, binding with high affinity: NMB exhibits a Ki = 0.052 nM for the BB1R and about 1000 times higher for the BB2R, whereasBB2R, respectively GRP exhibits [1, 2a]. Ki They = 0.19 are nM selective for the agonists BB2R and for theabout respective 1000 times receptors, higher bindingfor the BB1R with high[2]. Neitheraffinity: NMB NMB nor exhibits GRP abind Ki = to0.052 the BB nM3R, for and the although BB1R and its about endogenous 1000 times ligand higher has for not the been BB2R, disclosed, whereas severalGRP exhibits selective a Kipeptide= 0.19 and nM nonpeptide for the BB2R agonists and about have been 1000 disclosed times higher [2]. forOn the other BB1R hand, [2]. Neither a few peptideNMB nor antagonists GRP bind towith the a BB3R, diverse and degree although of itssele endogenousctivity have ligand also hasbeen not disclosed been disclosed, for the severalthree receptorsselective peptide[2,14,15]. and However, nonpeptide the agonistspoor oral have bioavailabi been disclosedlity, low [2 ].absorption, On the other rapid hand, degradation a few peptide by proteolyticantagonists enzymes with a diverse and degreeimmunogenic of selectivity profile have of also peptides been disclosed make nonpeptide for the three receptorsmolecules [2 ,more14,15]. desirableHowever, [16,17]. the poor Efforts oral bioavailability,in this direction low resulted absorption, in the rapiddiscover degradationy of second by generation proteolytic peptoids enzymes PD168368and immunogenic and PD176252 profile (Figure of peptides 1) [18], make along nonpeptide with a set of molecules analogs with more dive desirablerse substitutions [16,17]. E ff[19]orts thatin this exhibit direction an antagonist resulted in profile the discovery for the ofBB1R second and generation BB2R, with peptoids diverse PD168368 degrees of and selectivity. PD176252 Specifically,(Figure1)[ 18 PD168368], along with exhibits a set a of Ki analogs = 0.5 nM with for diverse the BB1R substitutions and 1700 [19 nM] that for exhibitthe BB2R, an antagonist whereas PD176252profile for exhibits the BB1R a Ki and = 0.5 BB2R, nM with for the diverse BB1R degrees and 170 of nM selectivity. for the BB2R Specifically, [2]. Furthermore, PD168368 the exhibits same a scaffoldKi = 0.5 was nM also for the later BB1R used and to 1700design nM ML-18, for the a BB2R,BB3R whereasselective PD176252antagonist exhibits[20], and a Kimore= 0.5 recently, nM for compoundsthe BB1R and AM-37 170 nM and for ST-36 the with BB2R diverse [2]. Furthermore, pharmacolo thegical same profile sca fftoold the was diverse also bombesin later used receptors to design [21].ML-18, However, a BB3R the selective similarity antagonist between [20 ],the and chemical more recently, structures compounds of these AM-37 compou andnds ST-36 is so with big diversethat it makespharmacological it intriguing profile to understand to the diverse the bombesinsubtle differences receptors that [21]. provide However, selectivity the similarity for the between different the receptors.chemical structuresIn the absence of these of ligand–receptor compounds is so3D big complex that it makesstructures, it intriguing rationalization to understand of the structure– the subtle activitydifferences needs that to providebe carried selectivity out indirectly, for the dialthoufferentgh receptors. the lack of In structural the absence diversity of ligand–receptor of the known 3D binderscomplex makes structures, it a difficult rationalization job. Complementary of the structure–activity to the structure–activity needs to be carried studies, out it indirectly,was pointed although a few yearsthe lack ago of in structural an interesting diversity report of the[22] knownthat Tyr220 binders in makesBB1R (corresponding it a difficult job. to Complementary Phe218 in BB2R) to can the explainstructure–activity the differential studies, behavior it was exhibited pointed aby few PD168368 years ago for in the an two interesting receptors. report [22] that Tyr220 in BB1R (corresponding to Phe218 in BB2R) can explain the differential behavior exhibited by PD168368 for the two receptors.

Figure 1. Chemical structures including the chirality of the asymmetric carbon (*) of the bombesin antagonists studied in the present work. PD168368 (1) and PD176252 (2). Figure 1. Chemical structures including the chirality of the asymmetric carbon (*) of the bombesin antagonistsDespite knowledge studied in accumulatedthe present work. in the PD168368 past to ( ascertain1) and PD176252 the potential (2). use of bombesin antagonists as therapeutic agents, advancement has been hindered because of the low diversity within the ligands available.Despite Moreover, knowledge PD168368 accumulated and PD176252 in the havepast been to ascertain reported tothe be potential potent agonists use of of thebombesin human antagonistsformyl-peptide as therapeutic receptors, agents, questioning advancement the interpretation has been thathindered their because reported of eff theects low can diversity be solely withinattributed the ligands to their activityavailable. as Moreover, BB1/BB2 antagonists PD168368 [and23]. PD176252 Only a few have compounds been reported devoid to of be a peptoidpotent agonistsscaffold of have the beenhuman disclosed formyl-peptide in the literature receptors, to date:questioning a BB1R the antagonist interpretation with athat dibenzodiazepine their reported effectsscaffold can [24 be], thesolely reported attributed antagonist to their of theactivity BB2 receptor as BB1/BB2 NSC-77427 antagonists [25] and [23]. Bantag-1, Only a few a peptidomimetic compounds devoiddesigned of a by peptoid an isostere scaffold replacement have been thatdisclosed exhibits in the a selective literature antagonist to date: aprofile BB1R antagonist for the BB3R with [26 a].

Pharmaceuticals 2020, 13, x FOR PEER REVIEW 3 of 16

dibenzodiazepine scaffold [24], the reported antagonist of the BB2 receptor NSC-77427 [25] and PharmaceuticalsBantag-1, a peptidomimetic2020, 13, 197 designed by an isostere replacement that exhibits a selective antagonist3 of 16 profile for the BB3R [26]. In order to have a better understanding of the therapeutic potential of Inbombesin order to antagonists, have a better there understanding is a need ofto the discov therapeuticer novel potential compounds of bombesin based on antagonists, different therechemical is a needstructures. to discover In this novel direction, compounds we report, based in the on diprffesenterent work, chemical the structures.results of a Inmolecular this direction, modeling we report, study inaimed the presentat discovering work, the novel results chemical of a molecular scaffolds modelingsuitable to study develop aimed novel at discovering bombesin antagonists. novel chemical For scathisff oldspurpose, suitable we toconstructed develop novel a 3D bombesin model of antagonists. the BB1 receptor For this by purpose, homology weconstructed modeling and a 3D docked model ofPD176252 the BB1 receptorin diverse by homologyposes. Analysis modeling of andthe dockedligand–receptor PD176252 complexes in diverse together poses. Analysis with known of the ligand–receptorstructure–activity complexes studies of together this family with of known compounds structure–activity permitted to studies define of a thispharmacophore family of compounds that was permittedsubsequently to define used for a pharmacophore in silico screening. that wasThe subsequentlyresults of the usedstudy for permitted in silico screening.the identification The results of a oflimited the study set of permitted small molecules the identification that were purchased of a limited an setd tested of small at molecules50 µM for its that capacity were purchased to antagonize and testedNMB at at the 50 µ BB1R.M for The its capacity results of to this antagonize study are NMB disclosed at the BB1R.in the Thepresent results report. of this study are disclosed in the present report. 2. Results and Discussion 2. ResultsFigure and 2 shows Discussion the result of the multiple alignment of the diverse sequences of the class-A rhodopsinFigure family2 shows used the in result the present of the work. multiple As can alignment be seen, ofthe the alignment diverse was sequences robust enough of the class-Ato show rhodopsinproperly aligned family useddiverse in theconserved present motifs, work. Asspecifical can bely, seen, a set the of alignment conserved was residues robust including enough to (in show the properlyBallesteros-Weinstein aligned diverse numbering conserved scheme motifs, [27]) specifically, N1.50 in a TM1; set of L2.46, conserved A2.47, residues D2.50 in including TM2; D/E3.49, (in the Ballesteros-WeinsteinR3.50, Y3.51 in TM3; W4.50 numbering in TM4; scheme F5.47, [ 27P5.50,]) N1.50 Y5.58 in in TM1; TM5; L2.46, F6.44, A2.47, W6.48, D2.50 P6.50 in in TM2; TM6; D /N7.49,E3.49, R3.50,P7.50, Y3.51Y7.53 inin TM3;TM7. W4.50Focusing in TM4; on the F5.47, sequences P5.50, Y5.58of thein NTS1 TM5; and F6.44, BB1 W6.48, receptors, P6.50 the in conserved TM6; N7.49, residue P7.50, Y7.53Y5.58 inin TM7.the former Focusing was on an the N5.58, sequences but this of difference the NTS1 anddid BB1not induce receptors, any the structural conserved consequence. residue Y5.58 In incontrast, the former Y7.53 was in the an N5.58,former but corresponded this difference to Y7.54 did not in inducethe latter. any A structural closer inspection consequence. of the In alignment contrast, Y7.53suggests in the that former this corresponded corresponded to to a Y7.54displacement in the latter. and not A closer an insertion inspection that of could the alignment induce a buldge suggests in thatTM7. this Other corresponded motifs such to as a D(E)RY displacement in TM3, and CWxP(Y/F/L) not an insertion in TM6 that or could the NPxxY induce in a buldgeTM7 were in TM7.well Otheraligned. motifs such as D(E)RY in TM3, CWxP(Y/F/L) in TM6 or the NPxxY in TM7 were well aligned.

Figure 2. Multiple sequence alignment of diverse GPCRs used in the present work (see text). TransmembraneFigure 2. Multiple segments sequence are coloredalignment in orangeof diverse and conservedGPCRs used residues in the are present in red work boxes. (see There text). is alsoTransmembrane a purple line indicatingsegments aare disulfide colored bridge. in orange and conserved residues are in red boxes. There is also a purple line indicating a disulfide bridge.

Pharmaceuticals 2020, 13, x FOR PEER REVIEW 4 of 16 Pharmaceuticals 2020, 13, 197 4 of 16 Figure 3 shows the evolution of the root-mean-square-deviation (rmsd) computed using the Cα of theFigure BB1 3receptor shows thestructure evolution along of thethe root-mean-square-deviationMD trajectory. A can be seen, (rmsd) it required computed more using than the 100 C nsα forof thethe BB1 structure receptor to structurehave the along structure the MD equilibrated trajectory. as A canhad bebeen seen, observed it required previously more than with 100 nsother for MD the structuresimulations to have of GPCRs the structure [28]. The equilibrated procedure as hadpermitte beend observed the construction previously of with a final other model MD simulations of the BB1 ofreceptor GPCRs that [28]. was The generated procedure as permitted the average the constructionstructure obtained of a final using model the oflast the 100 BB1 ns receptor of the molecular that was generateddynamics astrajectory. the average This structure structure obtained was subsequent using thely last minimized 100 ns of in the a moleculartwo-step procedure, dynamicstrajectory. using the Thissteepest structure descent was method subsequently with a minimized distance dependent in a two-step dielectric procedure, constant using theof 2 steepest. First, side descent chains method were withoptimized a distance with dependent the backbone dielectric atoms constant constrained of 2. First, sideto be chains subsequently were optimized released with thein backbonea second atomsminimization constrained step. to be subsequently released in a second minimization step.

Figure 3. Time evolution of the root mean square deviation (rmsd) of the bombesin BB1 receptor during Figure 3. Time evolution of the root mean square deviation (rmsd) of the bombesin BB1 receptor the refinement process. The rmsd was computed using the alpha carbons of the protein. during the refinement process. The rmsd was computed using the alpha carbons of the protein. In order to understand the molecular features that ligands need to bind onto the BB1 receptor, In order to understand the molecular features that ligands need to bind onto the BB1 receptor, we analyzed the complex of PD168368 and PD176252, respectively, in their prospective bound we analyzed the complex of PD168368 and PD176252, respectively, in their prospective bound conformations. For this purpose, we exploited the available structure–activity studies [18,19], the conformations. For this purpose, we exploited the available structure–activity studies [18,19], the available results from mutagenesis studies [22] as well as the small residue differences between the available results from mutagenesis studies [22] as well as the small residue differences between the BB1R and the BB2R at the orthosteric site. Accordingly, the constructed model of the BB1 receptor BB1R and the BB2R at the orthosteric site. Accordingly, the constructed model of the BB1 receptor was used for docking the antagonists in diverse conformations and orientations. For this purpose, was used for docking the antagonists in diverse conformations and orientations. For this purpose, we we followed the same protocol as described above. Multiple orientations were obtained that were followed the same protocol as described above. Multiple orientations were obtained that were rank rank ordered using the XP scoring function of GLIDE [29]. The diverse complexes were subsequently ordered using the XP scoring function of GLIDE [29]. The diverse complexes were subsequently energy minimized in vacuo with a distance dependent dielectric constant of 2, using the steepest energy minimized in vacuo with a distance dependent dielectric constant of 2, using the steepest descent method to allow the ligand to adapt to the environment. descent method to allow the ligand to adapt to the environment. Taking the asymmetric carbon of the ligands as reference, the structure of both compounds can Taking the asymmetric carbon of the ligands as reference, the structure of both compounds can be described as composed of three branches including a nitrophenylurea moiety, an indole moiety be described as composed of three branches including a nitrophenylurea moiety, an indole moiety and a 2-pryridinecyclohexane moiety. The diverse orientations and conformations found permitted and a 2-pryridinecyclohexane moiety. The diverse orientations and conformations found permitted the three legs to occupy diverse sites of the orthosteric site of the BB1 receptor and analyze diverse the three legs to occupy diverse sites of the orthosteric site of the BB1 receptor and analyze diverse ligand–receptor interactions. As can be expected, the orientations adopted by PD168368 and PD176252 ligand–receptor interactions. As can be expected, the orientations adopted by PD168368 and were comparable due to their size and the small structural differences between them. Actually, the PD176252 were comparable due to their size and the small structural differences between them. difference between the two molecules consisted of an extra methoxyl group at the pryridine moiety Actually, the difference between the two molecules consisted of an extra methoxyl group at the in PD176252, as shown in Figure1. However, this extra moiety provided PD176252 with an order of pryridine moiety in PD176252, as shown in Figure 1. However, this extra moiety provided PD176252 magnitude improved affinity to the BB2R that can be attributable to an additional favorable interaction with an order of magnitude improved affinity to the BB2R that can be attributable to an additional with the receptor. favorable interaction with the receptor.

Pharmaceuticals 2020, 13, x FOR PEER REVIEW 5 of 16 Pharmaceuticals 2020, 13, 197 5 of 16

Therefore, in order to assess the most likely binding mode of the ligands to the receptor, we analyzedTherefore, the diverse in order ligand–receptor to assess the interactions, most likely aimed binding at modethe identification of the ligands of key to the residues, receptor, and we checkedanalyzed their the conservation diverse ligand–receptor in the BB2R interactions,to understand aimed if a specific at the identificationinteraction could of key be residues,responsible and forchecked the observed their conservation differential binding in the BB2R affinity to understandof the compounds if a specific for the interaction two receptors. could This be responsible assumes thatfor differences the observed in the diff erentialfree energy binding of binding affinity be oftween the compounds PD168368 and for thePD176252 two receptors. are attributable This assumes only tothat the dienthalpicfferences component. in the free energy Obviously, of binding differences between in PD168368the affinity and of PD176252the two compounds are attributable could only also to bethe attributed enthalpic to component.the ability of Obviously,the differential differences chemical in group the affi tonity kick of out the a two bound compounds water molecule could alsofrom be theattributed receptor, to improving the ability the of affinity the diff erentialof a ligand chemical by an grouporder of to magnitude kick out a bound[30]. As water mentioned molecule before, from earlythe receptor,site-directed improving mutagenesis the affi nitystudies of a demonstrat ligand by aned order the involvement of magnitude of [30 Tyr220]. As mentioned of BB1 in before, the bindingearly site-directed of PD168368. mutagenesis Actually, when studies Tyr220 demonstrated is mutated the to involvement a phenylalanine, of Tyr220 binding of BB1 drops in the to binding that observedof PD168368. for the Actually, BB2R [22]. when This Tyr220 result issuggests mutated that to PD168368, a phenylalanine, and presumably binding drops PD176252, to that exhibit observed a criticalfor the interaction BB2R [22]. with This Tyr220 result suggestsin the BB1R that PD168368,bound complex and presumably that disappears PD176252, when exhibitthe residue a critical is mutatedinteraction to Phe220. with Tyr220 Since inthe the replacement BB1R bound of complexTyr for Phe that represents disappears a whenconservative the residue mutation, is mutated it can to bePhe220. concluded Since that the the replacement hydroxyl group of Tyr of for the Phe tyrosine represents is involved a conservative in a hydrogen mutation, bond it canwith be the concluded ligand. Thisthat represents the hydroxyl an groupimportant of the structural tyrosine isconstraint involved inthat a hydrogen permits us bond to discard with the a ligand. few poses This from represents the dockingan important study. Analysis structural of constraint the diverse that poses permits consiste usntly to discard showed a a fewsubset poses of them from with the dockingthe hydroxyl study. groupAnalysis of the of Tyr220 the diverse side poseschain pointing consistently towards showed the acenter subset of of the them aromatic with thering hydroxyl of the nitrobenzene group of the ringTyr220 of the side ligand, chain as pointing shown towardsin Figure the 4. centerInterestingly, of the aromaticthe interaction ring of between the nitrobenzene both moieties ring ofwas the reinforcedligand, as by shown a quadruple–quadruple in Figure4. Interestingly, interaction the interaction between between the aromatic both moietiesring of the was ligand reinforced and bythe a aromaticquadruple–quadruple ring of the tyrosine interaction side chain between adopting the aromatic a T-shape ring relative of the ligandorientation. and the This aromatic explains ring in part of the thetyrosine lower sideaffinity chain exhibited adopting by a PD168368 T-shape relative and PD176252 orientation. for Thisthe BB2R. explains In inthis part case, the although lower affi thenity quadrupole–quadrupoleexhibited by PD168368 andinteraction PD176252 between for the aromatic BB2R. In thisrings case, was although preserved the in quadrupole–quadrupole the complex with the BB2R,interaction the loss between of a hydrogen aromatic bond rings can was be preserved associated in with the complexan order with of magnitude the BB2R, drop the loss of affinity of a hydrogen [31]. Inbond addition, can be the associated nitro group with formed an order a hydrogen of magnitude bond drop with of His283 affinity (residue [31]. In conserved addition, thein the nitro BB2R), group reinforcingformed a hydrogenthe role played bond withby the His283 nitrophenyl (residue moiety conserved in the in affinity the BB2R), of the reinforcing two ligands, the roleas shown played in by Figurethe nitrophenyl 4. moiety in the affinity of the two ligands, as shown in Figure4.

Figure 4. Close-up of PD176252 in its prospective bound conformation to the bombesin BB1 receptor, Figureshowing 4. Close-up the interaction of PD176252 between in its its prospective nitrophenyl bound moiety conformation and diverse residuesto the bombesin BB1R including BB1 receptor, Tyr220 showingand His286 the interaction (see text). between its nitrophenyl moiety and diverse residues BB1R including Tyr220 and His286 (see text). The indole moiety in both ligands sat in a hydrophobic pocket defined by residues Phe181, Pro120 andThe Leu215 indole (Figure moiety5). in In both addition, ligands the sat indole in a moiety hydrophobic was oriented pocket indefined such a by way residues that it formedPhe181, a Pro120hydrogen and Leu215 bond with (Figure Glu178. 5). In All addition, these residues the indole were moiety conserved was oriented in the BB2R, in such so thesea way interactions that it formed help a tohydrogen explain bond part of with the Glu178. affinity ofAll the these ligand, residues but theywere do conserved not explain in the the BB2R, higher so a ffithesenity interactions observed by helpPD168368 to explain and part PD176252 of the affinity towards of the the BB1R. ligand, but they do not explain the higher affinity observed by PD168368 and PD176252 towards the BB1R.

Pharmaceuticals 2020, 13, 197 6 of 16 Pharmaceuticals 2020, 13, x FOR PEER REVIEW 6 of 16

Figure 5. Close-up of PD176252 in its prospective bound conformation to the bombesin BB1 receptor, Figure 5. Close-up of PD176252 in its prospective bound conformation to the bombesin BB1 receptor, showing the interaction between its indole moiety and residues Phe181, Pro120 and Leu215 (see text). showing the interaction between its indole moiety and residues Phe181, Pro120 and Leu215 (see text). To understand the difference in the binding affinity between PD168368 and PD176252 towards the To understand the difference in the binding affinity between PD168368 and PD176252 towards BB2R we should focus on the role of the methoxyl group attached to the 2-pyridinecyclohexane moiety. the BB2R we should focus on the role of the methoxyl group attached to the 2-pyridinecyclohexane As mentioned above, due to the similar size of the ligands, it can be hypothesized that PD168368 and moiety. As mentioned above, due to the similar size of the ligands, it can be hypothesized that PD176252 adopt a similar bound conformation to the BB1R. Analysis of the diverse poses with the PD168368 and PD176252 adopt a similar bound conformation to the BB1R. Analysis of the diverse nitrophenyl moiety accommodated in the proximity to Tyr220, revealed a subset of poses, where the poses with the nitrophenyl moiety accommodated in the proximity to Tyr220, revealed a subset of methoxypyridinyl group nicely sat in a pocketin which the side chain of Ser126 (residue conserved in poses, where the methoxypyridinyl group nicely sat in a pocketin which the side chain of Ser126 the BB2R) is located at a suitable distance to form a hydrogen bond with the oxygen of the methoxyl (residue conserved in the BB2R) is located at a suitable distance to form a hydrogen bond with the moiety (Figure6). In addition, the side chain of Arg310 (also conserved in BB2R) that formed a oxygen of the methoxyl moiety (Figure 6). In addition, the side chain of Arg310 (also conserved in hydrogen bond with Asp100 in the apo form, was in the position to form a hydrogen bond with the BB2R) that formed a hydrogen bond with Asp100 in the apo form, was in the position to form a center of the aromatic ring of the PD176252 pyridine ring. The presence of a hydrogen bond between hydrogen bond with the center of the aromatic ring of the PD176252 pyridine ring. The presence of a the ligand and Ser126 can explain the affinity difference between PD168368 and PD176252 towards hydrogen bond between the ligand and Ser126 can explain the affinity difference between PD168368 the BB2R. However, it does not explain that the affinity of both compounds to the BB1R remained and PD176252 towards the BB2R. However, it does not explain that the affinity of both compounds the same. This can be explained on the basis that the 2-pyridine moiety did not appear to bind in to the BB1R remained the same. This can be explained on the basis that the 2-pyridine moiety did not the same conformation in both ligands. In the case of PD168368, the pyridine ring was likely to be appear to bind in the same conformation in both ligands. In the case of PD168368, the pyridine ring accommodated close to Arg289, in such a way that its side chain formed a hydrogen bond with the was likely to be accommodated close to Arg289, in such a way that its side chain formed a hydrogen nitrogen of the heterocycle (Figure7). Since this residue was conserved in both receptors, this e ffect was bond with the nitrogen of the heterocycle (Figure 7). Since this residue was conserved in both expected to be found in both the BB1R and the BB2R and does not contribute to explaining the affinity receptors, this effect was expected to be found in both the BB1R and the BB2R and does not contribute difference of the compound for the two receptors. However, the position of the heterocycle permitted to explaining the affinity difference of the compound for the two receptors. However, the position of an additional interaction with Phe105 via a quadrupole–quadrupole interaction that was not found in the heterocycle permitted an additional interaction with Phe105 via a quadrupole–quadrupole the BB2R, since this residue was not conserved (it corresponded to Leu102 in BB2R). In addition, the interaction that was not found in the BB2R, since this residue was not conserved (it corresponded to cyclohexane group accessed a hydrophobic pocket formed by residues Met287 (Leu285 in BB2R), the Leu102 in BB2R). In addition, the cyclohexane group accessed a hydrophobic pocket formed by conserved Ile296 and the aliphatic chain of the conserved Lys210. These differences, together with the residues Met287 (Leu285 in BB2R), the conserved Ile296 and the aliphatic chain of the conserved differential interaction with Tyr220, can explain the affinity difference of the compounds for the two Lys210. These differences, together with the differential interaction with Tyr220, can explain the receptors. Accordingly, the fact that PD176252 exhibited the same affinity for BB1R and BB2R could affinity difference of the compounds for the two receptors. Accordingly, the fact that PD176252 be explained as a compensation of interactions, in such a way that the gain of a hydrogen bond with exhibited the same affinity for BB1R and BB2R could be explained as a compensation of interactions, Ser126 may compensate the loss of the quadrupole–quadrupole interaction with Phe105. in such a way that the gain of a hydrogen bond with Ser126 may compensate the loss of the quadrupole–quadrupole interaction with Phe105.

Pharmaceuticals 2020, 13, x FOR PEER REVIEW 7 of 16 Pharmaceuticals 2020, 13, 197 7 of 16 Pharmaceuticals 2020, 13, x FOR PEER REVIEW 7 of 16

Figure 6. Close-up of PD176252 in its prospective bound conformation to the bombesin BB1 receptor, Figure 6. Close-up of PD176252 in its prospective bound conformation to the bombesin BB1 receptor, Figureshowing 6. Close-upthe interaction of PD176252 between in itsits prospective phenoxyl-2-pyr boundidine conformation moiety interacting to the bombesin with diverse BB1 receptor, residues showing the interaction between its phenoxyl-2-pyridine moiety interacting with diverse residues showingincluding the Ser126, interaction Gln123 between and Arg310 its phenoxyl-2-pyr (see text). idine moiety interacting with diverse residues including Ser126, Gln123 and Arg310 (see text). including Ser126, Gln123 and Arg310 (see text).

Figure 7. Close-up of PD168368 in its prospective bound conformation to the bombesin BB1 receptor,

showingFigure the 7. Close-up interaction of betweenPD168368 its in phenoxyl-2-pyridine its prospective bound moiety conformation and diverse to residues the bombesin including BB1 Arg289 receptor, andFigureshowing Phe105 7. Close-up (seethe interaction text). of PD168368 between in its its prospective phenoxyl-2-p boundyridine conformation moiety and to thediverse bombesin residues BB1 includingreceptor, showingArg289 and the Phe105interaction (see text).between its phenoxyl-2-pyridine moiety and diverse residues including AccordingArg289 and to Phe105 the previous (see text). analysis, we can propose prospective conformations of PD168368 and PD176252According bound to the previous BB1R. Figure analysis,8 shows we can pictorially propose the prospective two compounds conformations superimposed of PD168368 when and boundPD176252According to the bound orthosteric to theto the previous site. BB1R. Inspection analysis,Figure 8 we ofshows Figurecan prpictopose8 oriallyshows prospective the two nitrophenyl compounds conformations group superimposed sittingof PD168368 close when toand Tyr220PD176252bound and to His283boundthe orthosteric andto the the BB1R. indolesite. FigureInspection moiety 8 sittingshows of Figu inpict areorially hydrophobic 8 shows the thetwo pocket nitrophenyl compounds defined groupsuperimposed by Phe181, sitting Pro120 close when to andboundTyr220 Leu215 to and the for His283 orthosteric both and ligands, the site. indole as Inspection described moiety of above.sitting Figu inre These a8 hydrophobicshows residues the nitrophenyl were pocket all defined conserved group by sittingPhe181, in the close BB2R, Pro120 to soTyr220and it is Leu215 not and expected His283 for both thatand ligands, theythe indole can as providedescribed moiety ansitting above. explanation in Th a hydrophobicese residues for the a ffi werepocketnity all di defined ffconservederences by between Phe181,in the BB2R, BB1RPro120 so andandit is BB2R. Leu215not expected On for the both otherthat ligands, they hand, can asthe provide described 2-pyridincyclohexane an above.explan ationThese for residues moiety the affinity were apparently differencesall conserved bound between in the a di BB1RBB2R,fferent and so conformationitBB2R. is not Onexpected the in theother that two they hand, ligands. can theprovide In 2-pyridincyclohexa the an case explan of PD168368,ation nefor moietythe the affinity pyridine apparently differences ring adopted bound between anin extendeda BB1R different and conformation,BB2R.conformation On the whereas inother the two inhand, the ligands. case the of 2-pyridincyclohexaIn the the PD176252, case of PD168368, thene corresponding moiety the pyridineapparently dihedral ring bound angle adopted was in an twisteda differentextended to 60conformation◦conformation,. This differential in whereas the conformation two in ligands. the case permitted In of thethe casePD176252, the of pryridine PD168368, the corresponding ring the of pyridine PD176252 dihedral ring to access adopted angle a hydrophobicwas an twistedextended to pocketconformation,60°. This and differential form whereas a hydrogen conformation in the bond case of with permitted the thePD176252, side the chainpryridine the ofcorresponding Ser126, ring of as PD176252 explained dihedral to above.accessangle was a Inhydrophobic contrast,twisted to PD16836860°.pocket This and differential adopted form aa conformationhydrogenconformation bond thatpermitted with permitted the the side pryridine the chain formation of ring Ser126, of of PD176252 aas hydrogen explained to access bond above. a with hydrophobic In Arg289 contrast, togetherpocketPD168368 withand adoptedform a quadrupole–quadrupole a hydrogena conformation bond withthat interaction permittedthe side withch thaine Phe105.formation of Ser126, of as a explained hydrogen above. bond withIn contrast, Arg289 PD168368together with adopted a quadrupole–quadrupole a conformation that permitted interaction th withe formation Phe105. of a hydrogen bond with Arg289 together with a quadrupole–quadrupole interaction with Phe105.

Pharmaceuticals 2020, 13, 197 8 of 16 Pharmaceuticals 2020, 13, x FOR PEER REVIEW 8 of 16

Figure 8. Superposition of PD168368 (cyan) and PD176252 (yellow) in their prospective bound Figureconformations, 8. Superposition respectively, of PD168368 to the BB1R. (cyan) and PD176252 (yellow) in their prospective bound conformations, respectively, to the BB1R. In order to further assess the feasibility of both prospective bound conformations, we checked theIn extent order that to further these modelsassess the explained feasibility the of availableboth prospective structure–activity bound conformations, results of this we familychecked of thecompounds extent that [18 ].these Analysis models of theexplained activity ofthe the availabl diversee compoundsstructure–activity showed results that when of this the nitrofamily group of compoundsof the nitrophenyl [18]. Analysis moiety of was the attached activity of in the position diverse 3 or compounds when it was showed substituted that when by a protonthe nitro accepting group ofchemical the nitrophenyl group such moiety as nitrile,was attached the compound in position preserved 3 or when its it activity. was substituted However, by when a proton the nitroaccepting group chemicalwas placed group in position such as nitrile, 2 or substituted the compound by a group preserved with lowerits activity. proton However, accepting when capability, the nitro the group affinity wasdropped placed at in leastposition one 2 order or substituted of magnitude. by a group The with same lower trends proton were accepting observed capability, regarding the affinity binding droppedaffinity forat least the BB2R. one order On the of othermagnitude. hand, substitutionsThe same trends on the were 2-pydirydine observed moietyregarding that the preserved binding a affinityproton for accepting the BB2R. center On inthe position other hand, 4 had substituti similar behavior.ons on the 2-pydirydine moiety that preserved a proton accepting center in position 4 had similar behavior. Proof of Concept Proof ofUsing Concept the prospective bound conformation of PD176252 onto the BB1R, we proceeded to identify thoseUsing interactions the prospective that appeared bound conformation to be key for of PD17 ligand–receptor6252 onto the recognition, BB1R, we proceeded aimed at to defining identify a thosepharmacophore interactions that that could appeared be used toas be a querykey for for ligand–receptor an in silico virtual recognition, screening. aimed Accordingly, at defining a simple a pharmacophorethree-point pharmacophore that could be wasused defined as a query as shownfor an in in silico Figure virtual9. The screening. pharmacophore Accordingly, was defined a simple as three-pointsimply as possiblepharmacophore to identify was hits defined with as chemical shown scain ffFigureoldsof 9. diverseThe pharmacophore profiles. The was pharmacophore defined as simplyinvolved as possible Tyr220 into theidentify form hits of awith proton chemical accepting scaffolds center, of diverse aimed atprofiles. discovering The pharmacophore BB1R selective involvedcompounds, Tyr220 since in thethe BB2Rform doesof a notproton have accepting this capability. center, Specifically, aimed at discovering the 3-point pharmacophoreBB1R selective compounds,included: (i) since a proton the BB2R accepting does not center have in this the directioncapability. of Specifically, the OH bond the of 3-point the hydroxyl pharmacophore moiety of included:Tyr220 side (i) a chain, proton located accepting at 2.5 center Å of thein the hydroxyl direction hydrogen; of the OH (ii) abond hydrophobic of the hydroxyl center locatedmoiety atof a Tyr220point definedside chain, by the located side chains at 2.5 ofÅ Pro200,of the hydroxyl Phe181 and hydrogen; Leu215; (iii)(ii) a a hydrophobic proton donor center locatedlocated inat thea pointplane defined defined by by the the side atoms chains of the of carboxyl Pro200, groupPhe181 of and Asp100 Leu215; at 3.0 (iii) Å froma proton the centerdonor of center the two located oxygens. in the planeThe defined three-point by the pharmacophore atoms of the wascarboxyl used asgrou a queryp of Asp100 for an in at silico 3.0 Å screening from the of center diverse of databasesthe two oxygens.using the Molecular Operating Environment (MOE) program [32]. More precisely, the query was defined in the form of three spheres with diverse radii to introduce some tolerance in each of the pharmacophoric points. Specifically, the proton acceptor and donor spheres were defined with a 0.12 nm radius, whereas the hydrophobic was defined within a 0.18 nm radius. Two databases of 3D structures of small molecules were used for the screening process including the leadlike database included in the MOE software containing around 650,000 commercially available compounds [32] and the leads-now subset of the ZINC database containing approximately 4,200,000 unique molecules downloaded in 2015 [33]. For each compound, in addition to its 3D structure, the database includes a set of conformations generated using a build-up procedure from systematic conformational searches of molecular fragments.

Pharmaceuticals 2020, 13, x FOR PEER REVIEW 9 of 16 Pharmaceuticals 2020, 13, 197 9 of 16

Figure 9. 3-point pharmacophore defined by the geometries of a few residues characterizing the Figuresterochemical 9. 3-point features pharmacophore involved indefined BB1R bindingby the (seegeometri text).es of a few residues characterizing the sterochemical features involved in BB1R binding (see text). After screening the 3D databases, a few hundred hits were identified and subjected to a diversity analysisThe three-point [34]. For this pharmacophore purpose, molecules was used were encodedas a query as bitfor strings an in usingsilico thescreening typed atomof diverse triangle databases(TAT) methodology using the Molecular in which atomsOperating are grouped Environm inent trios (MOE) based program on information [32]. More about precisely, their chemical the querynature was and defined mutual in distancethe form [of35 ].three In a spheres next step, with the diverse distance radii between to introduce bit strings some was tolerance computed in each using of thethe Tanimotopharmacophoric coefficient points. [36 ].Specifically, Finally, molecules the proton were acceptor grouped and indonor clusters spheres using were the defined Jarvis-Patrick with a 0.12algorithm nm radius, [37]. Thiswhereas procedure the hydrophobic permitted us was to selectdefined a subsetwithin of a compounds0.18 nm radius. preserving Two databases the diversity of 3Dof structures the initial of set, small showing molecules the diversity were used of chemicalfor the screening scaffolds. process Representative including molecules the leadlike of thedatabase diverse includedclusters in were the MOE selected software according containing to the suitabilityaround 650,000 of chemical commercially groups available responsible compounds for each of[32] the andpharmacophoric the leads-now subset points of and the checking ZINC database that molecules containing may approximately not suffer steric 4,200,000 clashes. unique A set of molecules about fifty downloadedselected compounds in 2015 [33]. were For docked each compound, onto the BB1R in addi usingtion theto its GLIDE 3D structure, software the [29 database] and ranked includes order a accordingset of conformations to the XP scoring generated function. using Thirteen a build- compoundsup procedure among from those systematic with the conformational best score were purchased (AppendixA) and tested at 50 µM for their capacity to displace the radioligand to the searchesPharmaceuticals of molecular 2020, fragments.13, x FOR PEER REVIEW 10 of 16 BB1R,After as screening explained the in 3D the databases, methods section a few hundred [38]. The hi chemicalts were identified structures and of subjected the six compounds to a diversity were found to displace the radioligand used in the binding assays by more than 15%, as listed in Table1. analysisPharmaceuticals [34].Table For 1. this 2020Listing ,purpose, 13, xof FOR small PEER molecules molecules REVIEW were identified encoded in the as in bit silico strings screening using (see the text). typed Column atom 2triangle shows 10 of 16 (TAT)These methodology resultstheir chemical yielded in whichstructure a success atoms and rate columns are of aboutgrouped 3 and 50%, 4 in the as trios displacements found based inother on ofinformation the cases corresponding [39,40 about]. Table BB1Rtheir1 also and chemical listsBB2R the natureresults and ofradioligands,Table mutual radioligand 1. Listing distance respectively, displacement of small[35]. Inmolecules (in a percentage)next experiments step, identified the at 50 fordistance in µM the (n in BB1R= between2).silico and screening BB2R,bit strings (see suggesting text).was computedColumn that the 2 shows using ligands theare Tanimoto BB1Rtheir selective. coefficient chemical structure [36]. Finally, and columns molecules 3 and 4 werethe displacements grouped in of clusters the corresponding using the BB1R Jarvis-Patrick and BB2R algorithm [37].radioligands, This procedure respectively, permitted (in percentage) us to sele atBB1R(Neuromendin 50ct µMa subset (n = 2). of compounds B preservingBB2R(Gastrin-Releasing the diversity Table 1. Listing of small molecules identified in the in silico screening (see text). Column 2 shows of the initialCompound# set, showing Chemical the diversity Structure of chemicalReceptor) scaffolds. Radioligand Representative Peptide molecules Receptor) of the Radioligand diverse their chemical structure and columns 3 and 4 the displacements of the corresponding BB1R and BB2R Displacement (%) Displacement (%) clustersradioligands, were selected respectively, according (in percentage)to the suitabilit at 50BB1R(NeuromendinµyM( ofn chemical= 2). groups B responsibleBB2R(Gastrin-Releasing for each of the Compound# Chemical Structure Receptor) Radioligand Peptide Receptor) Radioligand pharmacophoric points and checking that molecules may not suffer steric clashes. A set of about fifty selected compounds were docked onto the BB1RBB1R(Neuromendin usingDisplacement the GLIDE (%) B softwareBB2R(Gastrin-Releasing [29]Displacement and ranked Peptide(%) order accordingCompound# to the XP scoring Chemical function. Structure ThirteenReceptor) compounds Radioligand among those Receptor)with the Radioligandbest score were Displacement (%) Displacement (%) purchased (Appendix A) and tested at 50 µM for their capacity to displace the radioligand to the 1 19.7 0.0 BB1R, as explained in the methods section [38]. The chemical structures of the six compounds were found to displace the radioligand used in the binding assays by more than 15%, as listed in Table 1. 1 1 19.719.7 0.0 0.0 These results yielded a success rate of about 50%, as found in other cases [39,40]. Table 1 also lists the results of radioligand displacement experiments for the BB1R and BB2R, suggesting that the ligands are BB1R selective.

2 2 24.324.3 0.0 0.0

2 24.3 0.0

3 28.1 0.0

3 28.1 0.0

4 30.5 0.0

4 30.5 0.0

5 38.0 10.5

5 38.0 10.5

6 16.1 0.0

6 16.1 0.0

Pharmaceuticals 2020, 13, x FOR PEER REVIEW 10 of 16

Pharmaceuticals 2020, 13, x FOR PEER REVIEW 10 of 16 Table 1. Listing of small molecules identified in the in silico screening (see text). Column 2 shows their chemical structure and columns 3 and 4 the displacements of the corresponding BB1R and BB2R Table 1. Listing of small molecules identified in the in silico screening (see text). Column 2 shows Pharmaceuticalsradioligands, 2020 , respectively,13, x FOR PEER (in REVIEW percentage) at 50 µM (n = 2). 10 of 16 Pharmaceuticalstheir chemical 2020, 13 structure, x FOR PEER and REVIEWcolumns 3 and 4 the displacements of the corresponding BB1R and BB2R 10 of 16 radioligands,Table 1. Listing respectively, of small molecules(in percentage) identified at 50 µMin the (n =in 2). silico screening (see text). Column 2 shows Table 1. Listing of small molecules identifiedBB1R(Neuromendin in the in silico screening B (seeBB2R(Gastrin-Releasing text). Column 2 shows their chemical structure and columns 3 and 4 the displacements of the corresponding BB1R and BB2R Compound#their chemical Chemicalstructure and Structure columns 3 and Receptor)4 the displacements Radioligand of the correspondingPeptide Receptor) BB1R and Radioligand BB2R radioligands, respectively, (in percentage) BB1R(Neuromendinat Displacement50 µM (n = 2). (%) B BB2R(Gastrin-ReleasingDisplacement (%) Compound#radioligands, respectively,Chemical Structure (in percentage) atReceptor) 50 µM (n Radioligand = 2). Peptide Receptor) Radioligand BB1R(NeuromendinDisplacement (%) B BB2R(Gastrin-ReleasingDisplacement (%)

Compound# Chemical Structure BB1R(NeuromendinReceptor) Radioligand B PeptideBB2R(Gastrin-Releasing Receptor) Radioligand Compound# Chemical Structure Receptor) Radioligand Peptide Receptor) Radioligand Displacement (%) Displacement (%) Displacement (%) Displacement (%) 1 19.7 0.0

1 19.7 0.0

1 19.7 0.0 1 19.7 0.0

2 24.3 0.0

2 24.3 0.0

Pharmaceuticals 2020, 13, 197 10 of 16 2 24.3 0.0 2 24.3 0.0

Table 1. Cont.

BB1R(Neuromendin B BB2R(Gastrin-Releasing Peptide

Compound#3 Chemical Structure Receptor) Radioligand28.1 Receptor) Radioligand0.0 Displacement (%) Displacement (%)

3 28.1 0.0

3 28.1 0.0 3 28.1 0.0 3 28.1 0.0

4 30.5 0.0

4 30.5 0.0

4 30.5 0.0 4 30.5 0.0 4 30.5 0.0

5 38.0 10.5

5 38.0 10.5 5 38.0 10.5

5 38.0 10.5 5 38.0 10.5

6 16.1 0.0 6 16.1 0.0

6 16.1 0.0

6 16.1 0.0 6 16.1 0.0

In order to explain these results, we proceeded to dock the novel hits disclosed onto the BB1R following the same protocol as explained in the methods section. As an example, Figure 10 shows pictorially compound #5 in its prospective bound conformation to the BB1R. As can be seen, the triazine ring works as a scaffold with three branches: a phenoxyl moiety, a chlorophenyl moiety and an amine. Analysis of the prospective bound conformation showed that the oxygen of the methoxyl group fulfilled pharmacophoric point #1, exhibiting a hydrogen bond interaction with the hydroxyl moiety of Tyr220. Moreover, the chlorine atom sat at the hydrophobic site (pharmacophoric point #2) and the amine fulfilled pharmacophoric point #3 exhibiting a hydrogen bond with Asp300. In the case of the BB2R, these ligands could not attain the hydrogen bond with Tyr220, fulfilling only two pharmacophoric points, explaining that their affinity for the BB2R is much lower than for the BB1R. PharmaceuticalsPharmaceuticals 2020, 132020, x, FOR13, 197 PEER REVIEW 11 of 1611 of 16

FigureFigure 10. Pictorial 10. Pictorial view view of the of the proposed proposed binding binding mode mode of CompoundCompound # 5#5to to the the bombesin bombesin BB1 BB1 receptor. receptor. Finally, although the affinities exhibited by the hits presently disclosed is not very high, they representFinally, although a set of smallthe affinities molecules exhibited with high by diversitythe hits presently that can bedisclosed exploited is furthernot very to high, discover they novel representselective a set antagonists of small molecules for the BB1R. with high diversity that can be exploited further to discover novel selective3. Materials antagonists and for Methods the BB1R.

3. Materials3.1. Molecular and Methods Modeling

3.1. MolecularA crude Modeling model of the human bombesin BB1 receptor was constructed by homology modeling using the rat receptor NTS1 (PDB entry: 4GRV) as a template [41]. The template was selectedA crude formodel being of one the ofhuman the few bombesin receptors BB1 with rece a knownptor was crystallography constructed by structure homology located modeling in the same usingbranch the rat of neurotensin bombesin in receptor the GPCRs NTS1 phylogenetic (PDB entry: tree 4GRV) [28,42 as]. Sincea template the 4GRV [41]. structure The template corresponds was to selecteda fusion for being protein one of of NTS1 the few and receptors the T4 Lysozyme, with a known the template crystallography structure wasstructure edited located by removing in the the samecoordinates branch of ofbombesin the latter andin the joining GPCRs the segmentsphylogenetic of the tree ECL3 [28,42]. left at bothSince sides. the 4GRV In a subsequent structure step, correspondsthe sequences to a fusion of the protein BB1 and of the NTS1 NST1 and receptors the T4 Lysozyme, were aligned. the Thistemplate process structure is critical was for edited ensuring by the removingaccuracy the ofcoordinates the models of constructed the latter and by homologyjoining the modeling segments [ 43of ].the Since ECL3 in thisleft case,at both sequence sides. In identity a subsequentis low (~23%;step, the 43% sequences sequence of similarity), the BB1 and we undertook the NST1 areceptors multiple were sequence aligned. alignment This process to improve is its criticalquality for ensuring [44]. Accordingly, the accuracy we of selectedthe models a set constructed of 20 sequences by homology of diverse modeling GPCRs [43]. of theSince class-A in this with case,a sequence known crystallographic identity is low (~23%; structure 43% plus sequence the three similarity), bombesin we receptors undertook that a weremultiple aligned sequence using the alignmentCLUSTALW to improve software its quality (version [44]. 2.0.12) Accordingly, [45]. To we finalize selected the a alignmentset of 20 sequences process, of the diverse two sequences GPCRs of of theinterest class-A were with checked a known to ensurecrystallographic the alignment struct of diverseure plus known the three conserved bombesin residues receptors in the that class were A family alignedof GPCRs using the and CLUSTALW motifs such assoftware D(E)RY (version in TM3, 2.0. CWxP(Y12) [45]./F /ToL) infinalize TM6 orthe the alignment NPxxY in process, TM7, as the well as two sequencesthe location of of interest the conserved were checked disulfide to ensure bridge the between alignment ECL2 of and diverse TM3. known In a further conserved step, theresidues alignment in thewas class used A family to thread of GPCRs the sequence and motifs of the such BB1 as receptor D(E)RY onto in TM3, the backbone CWxP(Y/F/L) of the in template TM6 or the structure, NPxxY using in TM7,the as Molecular well as the Operating location of Environment the conserved (MOE) disulfide program bridge (version between 2019.01) ECL2 and [32]. TM3. Thirty In a models further were step,produced the alignment after thewas incorporation used to thread of alternativethe sequence sidechain of the BB1 conformations, receptor onto using the anbackbone extensive of rotamerthe templatelibrary structure, embedded using in MOE, the Molecular generated Operatin from a high-resolutiong Environment structural (MOE) prog database.ram (version Once hydrogens 2019.01) were [32]. addedThirty tomodels the models were usingproduced the protonate3D after the incorporation method [46], of each alternative of them wassidechain energy conformations, minimized using a usingcontact an extensive energy rotamer function library to relieve embedded any serious in MO stericE, generated strains. Models from a werehigh-resolution checked for structural inter-residue database.contacts Once as wellhydrogens as for backbonewere added conformations to the mode throughls using thethe Ramachandran protonate3D method map and [46], scored. each Finally,of themthe was model energy with minimized the lower using root-mean a contact square energy deviation function (rmsd) to inrelieve regard any to theserious average steric structure strains. with Modelsthe were highest checked score wasfor inter-residue selected as a crudecontacts model as well and as considered for backbone for refinement. conformations through the RamachandranRefinement map and of thescored. receptor Finally, was the carried model out with using the lower molecular root-mean dynamics square with deviation a ligand (rmsd) bound in in regardthe orthosteric to the average site. The structure presence with of a ligandthe highest provides score a more was eselectedfficient refinementas a crude of model the constructed and consideredmodel [for28]. refinement. Accordingly, the antagonist PD176252 (Figure1)[ 17] was docked in different conformations Refinement of the receptor was carried out using molecular dynamics with a ligand bound in the orthosteric site. The presence of a ligand provides a more efficient refinement of the constructed

Pharmaceuticals 2020, 13, 197 12 of 16 into the orthosteric site of the crude model using the GLIDE software (version 6.7) [29]. Multiple orientations were obtained that were rank ordered using the XP scoring function. The pose with the highest score was energy minimized in vacuo with a distance dependent dielectric constant of 2 using the steepest descent method to permit a reorientation of the ligand. Subsequently, the ligand–receptor complex was embedded in a lipid bilayer. Specifically, the protein was embedded in a box consisting of a 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) lipid and water molecules generated and equilibrated according to the procedure described previously [47]. The box had an initial size of 8.9 8.3 10.5 nm3 (XYZ), organized in such a way that the bilayer plane was oriented on the × × XY plane. The protein was placed in the center of the box, and the overlapping molecules were removed. Specifically, all water molecules with oxygen atoms closer than 0.40 nm to a nonhydrogen atom of the protein, as well as all lipid molecules with at least one atom closer than 0.25 nm to a nonhydrogen atom of the protein, were removed. This resulted in a final system containing 193 lipids and ca. 12,000 water molecules. Removal of these atoms introduced small voids between the protein and water or lipid molecules that disappeared during the first part of the molecular dynamics (MD) simulation, in which a progressive adjustment of the lipid bilayer and water molecules to the protein takes place. Next, 101 randomly selected water molecules were replaced by 47 sodium and 54 chloride ions, providing a neutral system with a concentration approximately 0.2 M in sodium chloride. This concentration is similar to that found in biological organisms, although they exhibit different intra- and extracellular ion concentrations. Then, the system was energy minimized to avoid steric classes using the steepest descent method and subjected to a 500 ns MD simulation at a constant pressure using the GROMACS package 4.6 [48]. Molecules were described using the all-atom OPLS force field [49] currently implemented in GROMACS, except for water molecules that were modeled using the TIP3P model [50]. The system was subjected to periodic boundary conditions in the three coordinate directions. The temperature was kept constant at 300 K using separate thermostats for the protein, water, ions and lipid molecules. The time constant for the thermostats was set to 0.1 ps, except for water, for which a smaller value of 0.01 ps was used. The pressure in the three coordinate directions was kept at 0.1 MPa by independent Berendsen barostats using a time constant of 1.0 ps. The equations of motion were integrated using the leapfrog algorithm with an integration step of 2 fs. All bonds involving hydrogen atoms within the protein and lipid molecules were kept frozen using the LINCS algorithm [51]. The bonds and the angle of water molecules were fixed using the analytical SETTLE method. Lennard–Jones interactions were computed using a cutoff of 1.0 nm. Electrostatic interactions were treated using the particle-mesh Ewald procedure [52].

3.2. Binding Assays BB1 antagonism assays were carried out following a protocol described elsewhere [42]. Specifically, human recombinant bombesin BB1 receptors expressed in CHO-K1 cells were used in modified HEPES-KOH buffer pH 7.4 (Thermo Scientific, Waltham, MA, USA). A 0.2 µg protein aliquot was 125 4 incubated with 0.1 nM [ I][Tyr ]-bombesin for 60 min at 25 ◦C. Nonspecific binding was estimated in the presence of 1 µM neuromedin B. Membranes were filtered and washed; the filters were then 125 4 counted to determine [ I][Tyr ]-bombesin (Kd = 0.13 nM) specifically bound. Hits were screened at 50 µM. Compound binding was calculated as the percentage of the inhibition of the binding of a radioactively labeled ligand.

4. Conclusions This paper reports the results of a modeling study aimed at shedding some light on the stereochemical requirements for small molecule binders to the BB1 bombesin receptor. For this purpose, we first constructed a 3D model of the bombesin BB1 receptor by homology modeling using the rat as a template. Then, the model was refined using molecular dynamics in a system composed by the receptor embedded in a bilayer of POCP lipids, water and sodium chloride. The MD simulation was carried out with the ligand PD176252 bound to the receptor in Pharmaceuticals 2020, 13, 197 13 of 16 its orthosteric site to accelerate the refinement process. After 500 ns sampling, the refined structure of the receptor was computed as the average of the diverse configurations sampled during the last 100 ns of the trajectory. Subsequently, the modeled 3D structure of the receptor was used to dock the antagonists PD168368 and PD176252 in its orthosteric site. Analysis of the complexes, guided by structure–activity and mutagenesis studies available, permitted us to propose prospective complexes of the bound conformation of each of the ligands to the BB1 receptor. The results of this study directly connect diverse pieces of information that were available in the literature and can be used as the basis for designing new experiments and small molecule ligands. As a proof of principle, we also carried out an in silico screening using a simple pharmacophore defined from the complex of PD176252 bound to the BB1R. Specifically, a three-point pharmacophore that involves a point exclusive for the BB1 receptor was used for this purpose. The in silico study permitted us to identify a set of small molecules with affinity for the BB1 receptor that were also disclosed. Interestingly, none of the molecules exhibited affinity for the BB2 receptor. The set of molecules have scaffolds of a diverse chemical nature that can be used as a starting point for the development of novel BB1 antagonists.

Author Contributions: Conceptualization, J.J.P.; methodology, P.G.-G.; calculations and formal analysis, B.R.; writing J.J.P. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Acknowledgments: The antagonist potency of the hits discovered in the in silico study were carried out by Eurofins under contract FR095-0015279 that is gratefully acknowledged. Conflicts of Interest: The authors declare no conflict of interest.

Appendix A Catalog Numbers and Suppliers of the Compounds Listed in Table1. Compound #1: AKOS000796000 (AKos GmbH, Stuttgart, Germany); Compound #2: Amb19684292 (Ambinter c/o Greenpharma SAS, Orleans, France); Compound #3: CB358 (Menai Organics Ltd., Bangor, UK); Compound #4: Amb1220897 (Ambinter c/o Greenpharma SAS, Orleans, France); Compound #5: Amb3992353 (Ambinter c/o Greenpharma SAS, Orleans, France); Compound #6: Amb11089933 (Ambinter c/o Greenpharma SAS, Orleans, France).

References

1. Jensen, R.T.; Moody, T.W. Bombesin Peptides (Cancer). In Hand-Book of Biologically Active Peptides; Kastin, A.J., Ed.; Elsevier: Amsterdam, The Netherlands, 2013; pp. 506–511. 2. Ramos-Alvarez, I.; Moreno, P.; Mantey, S.A.; Nakamura, T.; Nuche-Berenguer, B.; Moody, T.W.; Coy, D.H.; Jensen, R.T. Insights into bombesin receptors and ligands: Highlighting recent advances. Peptides 2015, 72, 128–144. [CrossRef] 3. Weber, H.C. Regulation and signaling of human bombesin receptors and their biological effects. Curr. Opin. Endocrinol. Diabetes Obes. 2009, 16, 66–71. [CrossRef][PubMed] 4. Flores, D.G.; De Farias, C.B.; Leites, J.; De Oliveira, M.S.; Lima, R.C.; Tamajusuku, A.S.; Leone, L.P.D.; Meurer, L.; Brunetto, A.L.; Schwartsmann, G.; et al. Gastrin releasing peptide receptors regulate proliferation of C6 glioma cells through a phosphatidylinositol 3-kinase-dependent mechanism. Curr. Neurovasc. Res. 2008, 5, 99–105. [CrossRef][PubMed] 5. Moody, T.W.; Moreno, P.; Jensen, R.T. as lung cancer growth factors. Peptides 2015, 72, 106–111. [CrossRef][PubMed] 6. Benya, R.V.; Kusui, T.; Pradhan, T.K.; Battey, J.F.; Jensen, R.T. Expression and characterization of cloned human bombesin receptors. Mol. Pharmacol. 1995, 47, 10–20. [PubMed] 7. Fathi, Z.; Corjay, M.H.; Shapira, H.; Wada, E.; Benya, R.; Jensen, R.; Viallet, J.; Sausville, E.A.; Battey, J.F. BRS-3: A novel bombesin receptor subtype selectively expressed in testis and lungcarcinoma cells. J. Biol. Chem. 1993, 268, 5979–8594. [PubMed] Pharmaceuticals 2020, 13, 197 14 of 16

8. Jensen, R.T.; Battey, J.F.; Spindel, E.R.; Beny, R.V. International Union of Pharmacology. LXVIII. Mammalian Bombesin Receptors: Nomenclature, Distribution, Pharmacology, Signaling, and Functions in Normal and Disease States. Pharm. Rev. 2008, 60, 1–42. [CrossRef] 9. Moreno, P.; Ramos-Alvarez, I.; Moody, T.W.; Jensen, R.T. Bombesin related peptides/receptors and their promising therapeutic roles in cancer imaging; targeting and treatment. Expert Opin. Ther. Targets 2016, 20, 1055–1073. [CrossRef] 10. Park, H.J.; Kim, S.R.; Kim, M.K.; Choi, K.S.; Jang, H.O.; Yun, I.; Bae, M.-K. antagonist suppresses tumor angiogenesis and tumor growth in vitro and in vivo. Cancer Lett. 2011, 312, 117–127. [CrossRef] 11. Gonzalez, N.; Moreno, P.; Jensen, R.T. Bombesin receptor subtype 3 as a potential target for obesity and diabetes. Expert Opin. Ther. Targets 2015, 19, 1153–1170. [CrossRef] 12. Zhou, S.; Nissao, E.; Jackson, I.L.; Leong, W.; Dancy, L.; Cuttitta, F.; Vujaskovic, Z.; Sunday, M.E. Radiation-Induced Lung Injury Is Mitigated by Blockade of Gastrin-Releasing Peptide. Am. J. Pathol. 2013, 182, 1248–1254. [CrossRef][PubMed] 13. Ehling, S.; Fukuyama, T.; Ko, M.C.; Olivry, T.; Bäumer, W. Neuromedin B Induces Acute Itch in Mice via the Activation of Peripheral Sensory Neurons. Acta Derm. Venereol. 2019, 99, 587–593. [CrossRef][PubMed] 14. Cristau, M.; Devin, C.; Oiry, C.; Chaloin, O.; Amblard, M.; Bernad, N.; Heitz, A.; Fehrentz, J.A.; Martinez, J. Synthesis and Biological Evaluation of Bombesin Constrained Analogues. J. Med. Chem. 2000, 43, 2356–2361. [CrossRef][PubMed] 15. Gonzalez, N.; Mantey, S.A.; Pradhan, T.K.; Sancho, V.; Moody, T.W.; Coy, D.H.; Jensen, R.T. Characterization of putative GRP- and NMB-receptor antagonist’s interaction with human receptors. Peptides 2009, 30, 1473–1486. [CrossRef] 16. Perez, J.J.; Corcho, F.; Llorens, O. Molecular modeling in the design of peptidomimetics and peptide surrogates. Curr. Med. Chem. 2002, 9, 2209–2229. [CrossRef] 17. Perez, J.J. Designing Peptidomimetics. Curr. Top. Med. Chem. 2018, 18, 566–590. [CrossRef] 18. Ashwood, V.; Brownhill, V.; Higginbottom, M.; Horwell, D.C.; Hughes, J.; Lewthwaite, R.A.; McKnight, A.T.; Pinnock, R.D.; Pritchard, M.C.; Suman-Chauhan, N.; et al. PD 176252-The First High Affinity Non-peptide Gastrin-Releasing Peptide (BB2) Receptor Antagonist. Bioorg. Med. Chem. Lett. 1998, 8, 2589–2594. [CrossRef] 19. Carrieri, A.; Lacivita, E.; Belviso, B.D.; Caliandro, R.; Mastrorilli, P.; Gallo, V.; Niso, M.; Leopoldo, M. Structural determinants in the binding of BB2 receptor ligands: In silico, x-ray and NMR studies in PD176252 analogues. Curr. Top. Med. Chem. 2017, 17, 1599–1610. [CrossRef] 20. Moody, T.W.; Mantey, S.A.; Moreno, P.; Nakamura, T.; Lacivita, E.; Leopoldo, M.; Jensen, R.T. ML-18 is a non-peptide bombesin receptor subtype-3 antagonist which inhibits lung cancer growth. Peptides 2015, 64, 55–61. [CrossRef] 21. Moody, T.W.; Tashakkori, N.; Mantey, S.A.; Moreno, P.; Ramos-Alvarez, I.; Leopoldo, M.; Jensen, R.T. AM-37 and ST-36 are Small Molecule Bombesin Receptor Antagonists. Front. Endocrinol. 2017, 8, 176. [CrossRef] 22. Tokita, K.; Hocart, S.J.; Katsuno, T.; Mantey,S.A.; Coy,D.H.; Jensen, R.T. Tyrosine 220 in the 5th Transmembrane Domain of the Neuromedin B Receptor Is Critical for the High Selectivity of the Peptoid Antagonist PD168368. J. Biol. Chem. 2001, 276, 495–504. [CrossRef][PubMed] 23. Schepetkin, I.A.; Kirpotina, L.N.; Khlebnikov, A.I.; Jutila, M.A.; Quinn, M.T. Gastrin-Releasing Peptide/Neuromedin B Receptor Antagonists PD176252, PD168368, and Related Analogs Are Potent Agonists of Human Formyl-Peptide Receptors. Mol. Pharmacol. 2011, 79, 77–90. [CrossRef][PubMed] 24. Fu, J.; Shuttleworth, S.J.; Connors, R.V.; Chai, A.; Coward, P. Discovery and optimization of a novel Neuromedin B receptor antagonist. Bioorg. Med. Chem. Lett. 2009, 19, 4264–4267. [CrossRef] 25. Martinez, A.; Zudaire, E.; Julian, M.; Moody, T.W.; Cuttitta, F. Gastrin-releasing peptide (GRP) induces angiogenesis and the specific GRP blocker 77427 inhibits tumor growth in vitro and in vivo. Oncogene 2005, 24, 4106–4113. [CrossRef][PubMed] 26. Moreno, P.; Mantey, S.A.; Nuche-Berenguer, B.; Reitman, M.L.; Gonzalez, N.; Coy, D.H.; Jensen, R.T. Comparative pharmacology of bombesin receptor subtype-3 nonpeptide agonist MK-5046, a universal peptide agonist; and peptide antagonist Bantag-1 for human bombesin receptors. J. Pharmacol. Exp. Ther. 2013, 347, 110–116. [CrossRef][PubMed] Pharmaceuticals 2020, 13, 197 15 of 16

27. Ballesteros, J.A.; Weinstein, H. Integrated methods for the construction of three-dimensional models and computational probing of structure-function relations in G protein-coupled receptors. Methods Neurosci. 1995, 25, 366–428. 28. Lupala, C.S.; Rasaifar, B.; Gomez-Gutierrez, P.; Perez, J.J. Using Molecular Dynamics for the refinement of atomistic models of GPCRs by homology modeling. J. Biomol. Struct. Dyn. 2018, 36, 2436–2448. [CrossRef] 29. Friesner, R.A.; Banks, J.L.; Murphy, R.B.; Halgren, T.A.; Klicic, J.J.; Mainz, D.T.; Repasky, M.P.; Knoll, E.H.; Shaw, D.E.; Shelley, M.; et al. Glide: A New Approach for Rapid Accurate Docking and Scoring. 1. Method and Assessment of Docking Accuracy. J. Med. Chem. 2004, 47, 1739–1749. [CrossRef] 30. Venkatakrishnana, A.J.; Maa, A.K.; Fonseca, R.; Latorraca, N.R.; Kelly, B.; Betz, R.M.; Asawa, C.; Kobilka, B.K.; Dror, R.O. Diverse GPCRs exhibit conserved water networks for stabilization and activation. Proc. Natl. Acad. Sci. USA 2019, 116, 3288–3293. [CrossRef] 31. Bissantz, C.; Kuhn, B.; Stahl, M. A Medicinal Chemist’s Guide to Molecular Interactions. J. Med. Chem. 2010, 53, 5061–5084. 32. Molecular Operating Environment (MOE), Version 2019.01; Chemical Computing Group UCL: Montreal, QC, Canada, 2020. 33. Sterling, T.; Irwin, J.J. ZINC 15—Ligand Discovery for Everyone. J. Chem. Inf. Model. 2015, 55, 2324–2337. [CrossRef][PubMed] 34. Perez, J.J. Managing molecular diversity. Chem. Soc. Rev. 2005, 34, 143–152. [CrossRef][PubMed] 35. Bender, A.; Jenkins, J.L.; Scheiber, J.; Chetan, S.; Sukuru, K.; Glick, M.; Davies, J.W. How Similar Are Similarity Searching Methods? A Principal Component Analysis of Molecular Descriptor Space. J. Chem. Inf. Model. 2009, 49, 108–119. [CrossRef][PubMed] 36. Lipkus, A.H. A proof of the triangular inequality for the Tanimoto distance. J. Math. Chem. 1999, 26, 263–265. [CrossRef] 37. Jarvis, R.A.; Patrick, E.A. Clustering Using a Similarity Measure Based on Shared Near Neighbors. IEEE Trans. Comp. 1973, 22, 1025–1034. [CrossRef] 38. Ohki-Hamazaki, H.; Iwabuchi, M.; Maekawa, F. Development and function of bombesin-like peptides and their receptors. Int. J. Dev. Biol. 2005, 49, 293–300. [CrossRef] 39. Krishna, S.; Singh, D.K.; Meena, S.; Datta, D.; Siddiqi, M.I.; Banerjee, D. Pharmacophore-Based Screening and Identification of Novel Human Ligase I Inhibitors with Potential Anticancer Activity. J. Chem. Inf. Model. 2014, 54, 781–792. [CrossRef] 40. Gomez-Gutierrez, P.; Campos, P.M.; Perez, J.J. Identification of a Novel Inhibitory Allosteric Site of MAP Kinases. PLoS ONE 2016, 11, e0167379. [CrossRef] 41. Noinaj, N.; White, J.F.; Shibata, Y.; Love, J.; Kloss, B.; Xu, F.; Gvozdenovic-Jeremic, J.; Shah, P.; Shiloach, J.; Tate, C.G.; et al. The crystal structure of the neurotensin receptor NTS1 in complex with neurotensin (8–13). Nature 2012, 490, 508–513. 42. Costanzi, S.; Skorski, M.; Deplano, A.; Habermehl, B.; Mendoza, M.; Wang, K.; Biederman, M.; Dawson, J.; Gao, J. Homology modeling of a Class A GPCR in the inactive conformation: A quantitative analysis of the correlation between model/template sequence identity and model accuracy. J. Mol. Graph. Model. 2016, 70, 140–152. [CrossRef] 43. Nayeem, A.; Sitkoff, D.; Krystek, S., Jr. A comparative study of available software for high-accuracy homology modeling: From sequence alignments to structural models. Protein Sci. 2006, 15, 808–824. [CrossRef] [PubMed] 44. Cavasotto, C.N.; Palomba, D. Expanding the horizons of G protein-coupled receptor structure-based ligand discovery and optimization using homology models. Chem. Commun. 2015, 51, 13576–13594. [CrossRef] [PubMed] 45. Larkin, M.A.; Blackshields, G.; Brown, N.P.; Chenna, R.; McGettigan, P.A.; McWilliam, H.; Valentin, F.; Wallace, I.M.; Wilm, A.; Lopez, R.; et al. Clustal W and Clustal X version 2.0. Bioinformatics 2007, 23, 2947–2948. [CrossRef][PubMed] 46. Labute, P. Protonate3D: Assignment of ionization states and hydrogen coordinates to macromolecular structures. Proteins 2009, 75, 187–205. [CrossRef][PubMed] 47. Cordomi, A.; Edholm, O.; Perez, J.J. Effect of different treatments of long-range interactions and sampling conditions in molecular dynamic simulations of embedded in a dipalmitoyl phosphatidylcholine bilayer. J. Comput. Chem. 2007, 28, 1017–1030. [CrossRef] Pharmaceuticals 2020, 13, 197 16 of 16

48. Van Der Spoel, D.; Lindahl, E.; Hess, B.; Groenhof, G.; Mark, A.E.; Berendsen, H.J. GROMACS: Fast flexible and free. J. Comput. Chem. 2005, 26, 1701–1718. [CrossRef] 49. Kaminski, G.; Friesner, R.A.; Tirado-Rives, J.; Jorgensen, W.L. Evaluation and Reparametrization of the OPLS-AA Force Field for Proteins via Comparison with Accurate Quantum Chemical Calculations on Peptides. J. Phys. Chem. B 2001, 105, 6474–6487. [CrossRef] 50. Jorgensen, W.L.; Chandrasekhar, J.; Madura, J.D.; Impey, R.W.; Klein, M.L. Comparison of simple potential functions for simulating liquid water. J. Chem. Phys. 1983, 79, 926. [CrossRef] 51. Miyamoto, S.; Kollman, P.A. Settle: An analytical version of the SHAKE and RATTLE algorithm for rigid water models. J. Comput. Chem. 1992, 13, 952–962. [CrossRef] 52. Darden, T.; York, D.; Pedersen, L. Particle mesh Ewald: An N log(N) method for Ewald sums in large systems. · J. Chem. Phys. 1993, 98, 10089. [CrossRef]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).