Structure–Signal Relationships of the δ-Opioid-Receptor (DOR)-Selective Agonist KNT-127—Part II: Quinoline Ring Modifications for Enhanced G-Protein Signaling and Reduced β-Arrestin Recruitment

Abstract

The δ-opioid receptor (DOR) continues to attract attention as a therapeutic target for the development of safer analgesics due to its ability to mediate pain relief with a lower risk of adverse effects compared to the μ-opioid receptor (MOR). Building upon our previous findings on KNT-127, a DOR-selective agonist with a morphinan scaffold, this study further explores the structure–signal relationships between quinoline ring modifications and the signaling bias toward Gi-protein activation while minimizing β-arrestin-2 recruitment. Our findings highlight the critical role of the 5′-position in modulating signaling bias. Bulky hydrophobic substituents, such as isopropoxy and cyclohexanoxy groups, effectively reduce β-arrestin-2 recruitment without compromising DOR binding affinity or Gi-protein activation. Molecular-docking and molecular dynamics simulations provided mechanistic insights, showing that these modifications change ligand interactions with the V281^6.55-W284^6.58-L300^7.35 sub-pocket, thus selectively favoring Gi-protein signaling. These insights clarify the key interactions for the signaling bias in DOR agonists, offering a new framework for the design of DOR-targeted therapies with an improved therapeutic profile.

Introduction

Opioid receptors (μ, MOR; δ, DOR; κ, KOR) are members of the G-protein-coupled receptor (GPCR) family. They are widely expressed in the nervous system and primarily mediate the inhibition of neuronal activity through Gi/o protein activation. MOR, the primary target of opioid analgesics such as morphine and fentanyl, induces strong analgesic effects. However, these drugs also exhibit serious side effects such as dependence, respiratory depression, constipation, and tolerance.^1,2) In contrast, DOR activation provides analgesia without the adverse effects associated with MOR.^3,4) Although DOR agonists exhibit lower analgesic efficacy in acute pain than MOR agonists, they have significant clinical expectations for chronic inflammatory and neuropathic pain as well as migraines.^5–7) Additionally, DOR is highly expressed in the limbic system, a region that is critical for emotional regulation, and DOR knockout mice have demonstrated depressive and anxiety-like behavior, highlighting the potential of DOR agonists in antidepressant and anxiolytic therapy.^8–10) Furthermore, DOR activation has shown neuroprotective effects, such as improved motor function in cases of Parkinsonian symptoms related to striatal dysfunction, and it also confers cardiovascular protection.^11–13) Given these benefits, DOR agonists hold significant potential for the development of novel therapeutic strategies across a range of diseases.

Despite this promising therapeutic potential, DOR agonists have yet to reach clinical use.¹⁴⁾ One main obstacle is the convulsion effects observed with certain diarylmethyl piperazine-type DOR agonists, such as SNC80 and BW373U86, at therapeutic doses^15,16) (Fig. 1A). Moreover, DOR agonists often lead to receptor internalization, resulting in unstable efficacy,¹⁷⁾ and have been reported to exhibit side effects including increased tolerance and ethanol consumption. While recent selective DOR agonists, such as AZD2327, ADL5747, and ADL5859, have entered clinical trials for indications including depression and inflammatory pain, concerns remain due to their structural similarity to SNC80. Independent pioneering studies by Kieffer and colleagues and van Rijn and colleagues revealed that the aforementioned side effects are linked to the recruitment of β-arrestin-2 (βarr2), which is associated with GPCR desensitization and intracellular trafficking. Studies using βarr2 knockout mice indicate that DOR agonists that induce less internalization of the receptor are less likely to cause desensitization and acute analgesic tolerance, and may also reduce convulsion effects, although the precise mechanism behind the convulsion susceptibility remains unclear.^18–20) Furthermore, β-arrestin signaling has been associated with DOR-mediated increases in alcohol consumption.²¹⁾ Multiple in vitro studies have demonstrated that diarylpiperazine derivatives, including SNC80, ARM390, and DPI-287, exhibit high β-arrestin recruitment efficacy despite their low potency.^21,22) In contrast, morphinan-based ligands, such as (−)-TAN-67 and KNT-127, display reduced β-arrestin recruitment potency and efficacy^21,23) (Fig. 1B). These findings suggest that structural differences among DOR agonists significantly impact β-arrestin recruitment activity; however, the impact of structural changes on this signal bias is still not completely understood.

Fig. 1. Chemical Structures of Diarylmethyl Piperazine- and Piperidine-Type (A), Morphinan-Type (B), and New Chemotype (C) DOR-Selective Agonists, and Structure–Signal Relationship Studies on KNT-127 (D)

In light of these findings, efforts have been made to develop DOR agonists that do not recruit β-arrestin. For example, van Rijn and colleagues have identified novel DOR-selective agonists with low βarr2 recruitment activity that feature a 1,3,8-triazaspiro [4.5]decan-2,4-dione²⁴⁾ or 3-thiophenyl-1H-pyrazole²⁵⁾ core through high-throughput screening of a relatively small commercial chemical library (Fig. 1C). Cheng and colleagues have focused on the amino-acid residues involved in the signaling bias toward Gi-protein activation, particularly N131^3.35 within the sodium-binding pocket, based on the cryo-electron microscopy structure of DOR with the piperidine-type agonist ADL5859.²⁶⁾ By optimizing the N-substituent of ADL5859, they identified an N-n-Bu derivative, ADL06, that exhibits attenuated βarr2 recruitment. Notably, ADL06 demonstrated strong analgesic effects at 29 mg/kg in a mouse model of complete Freund’s adjuvant-induced inflammatory pain without inducing seizures. While attenuating βarr2 signaling has proven effective in mitigating side effects, the efficacy of βarr2 recruitment remains approximately 40% for these ligands compared to reference ligands such as leucine-enkephalin (Leu-Enk) and ADL5859, leaving room for understanding the ligand–receptor interactions for further reducing βarr2 recruitment efficacy. Recently, we have investigated the structure–signal relationship of KNT-127, focusing on its quinoline ring²⁷⁾ (Fig. 1D). Our findings indicated that the morphinan scaffold tends to exhibit minimal βarr2 recruitment activity, while the introduction of a bicyclic aromatic quinoline ring enhances this activity.

Inspired by these results, in this study, we conducted further structure–signal-relationship studies centered on the substituents on the quinoline ring. We successfully reduced βarr2 recruitment efficacy to <20% of that observed with methionine-Enk while maintaining high Gi-protein activation by targeting interactions between the quinoline ring and key amino-acid residues in the receptor pocket, particularly L300^7.35. These findings point toward a promising strategy for the design of G-protein-signaling-selective DOR agonists.

Results and Discussion

In Vitro Pharmacology and Structure–Signal-Relationship Studies of Methoxy-Substituted Derivatives

Based on the single-crystal X-ray diffraction structure of the DOR antagonist naltrindole (NTI) complexed with DOR, we predicted that the quinoline ring of KNT-127 would occupy the same TM6-TM7 sub-pocket (V281^6.55-W284^6.58-L300^7.35) as the indole ring of NTI.²⁸⁾ Our recent studies have shown that removing 1 or 2 aromatic rings from the quinoline moiety of KNT-127 reduces the potency of Gi-protein dissociation activity while maintaining efficacy, but markedly decreases both the potency and efficacy of βarr2 recruitment.²⁷⁾ These findings suggest that interactions between the quinoline ring and the V281^6.55-W284^6.58-L300^7.35 sub-pocket play a critical role not only in Gi-protein activation but also in βarr2 recruitment. To investigate the influence of substituents at different positions of the quinoline ring on binding affinity and signaling bias, the methoxy group was chosen as a substituent for this study due to its optimal size, balanced lipophilicity, and hydrophilicity, which could potentially enhance interactions with the receptor binding pocket. Previous research by Ida et al. has explored the affinity of KNT-127 analogs with substituents at the 5′–8′ positions on the quinoline ring for DOR including methoxy-substituted derivatives, although the impact of these substitutions on G-protein activation and β-arrestin recruitment has not yet been investigated.²⁹⁾ With the aim of assessing how ligand interactions with the V281^6.55-W284^6.58-L300^7.35 sub-pocket influence binding affinity and signal transduction, we synthesized KNT-127 derivatives with methoxy groups at the 5′-, 6′-, 7′-, or 8′-position of the quinoline ring and evaluated their pharmacological activity (Fig. 1D).

Synthesis

Following previously reported synthetic methods,²⁹⁾ 3a–3d were synthesized by demethylation of morphinan 5^27,30) using BBr₃, followed by Friedländer quinoline synthesis under acidic conditions with 2-aminobenzaldehyde derivatives (3a: 80%; 3b: 76%; 3c: 56%; 3d: 93%; Chart 1). The structures of 3a–3d were assigned based on two-dimensional (2D)-NMR-based heteronuclear multiple-bond correlation (HMBC) analysis, which verified the presence of the quinoline ring on the morphinan C6–C7 position and the desired substituents at the 5′–8′ positions on the benzene ring (Supplementary Figs. S8–S11).

Chart 1. 

(a) BBr₃, CH₂Cl₂, 0°C (6: 55%; 4a: 67%); (b) 2-Aminobenzaldehyde derivatives, MeSO₃H, EtOH, reflux (3a: 80%; 3b: 76%; 3c: 56%; 3d: 93%; 4b: 71%; 4c: 79%; 4d: 61%; 4e: 52%); (c) 2-Amino-6-bromobenzaldehyde, 2% KOH/EtOH, microwave irradiation, 110°C (7: 98%); (d) 2-Isopropenyl-4,4,5,5-tetramethyl-1,3,2-dioxaborolane, Pd(PPh₃)₄, Cs₂CO₃, 1,4-dioxane, reflux; (e) Pd(OH)₂, H₂, 1,4-Dioxane, r.t. (8: 81% over 2 steps).

Binding Assay

To investigate the impact of methoxy substitution of KNT-127 at various positions of the quinoline ring on receptor affinity, we performed competitive binding assays using radiolabeled ligands across MOR, DOR, and KOR (Table 1 and Supplementary Fig. S1). Compared to KNT-127 (K_i = 0.36 nM; MOR/DOR = 171; KOR/DOR = 231), the 5′-methoxy derivative 3a displayed a 3.0-fold increase in DOR affinity (K_i = 0.12 nM) and high receptor selectivity toward MOR (MOR/DOR = 350), but slightly reduced receptor selectivity toward KOR (KOR/DOR = 127). The 6′-methoxy derivative 3b showed a 2.2-fold reduction in DOR affinity (K_i = 0.79 nM) and decreased receptor selectivity (MOR/DOR = 50; KOR/DOR = 69). In contrast,the 7′-methoxy derivative 3c exhibited only a 1.7-fold reduction in DOR affinity (K_i = 0.64 nM) while maintaining high receptor selectivity (MOR/DOR = 177; KOR/DOR = 453). The 8′-methoxy derivative 3d retained the DOR affinity (K_i = 0.34 nM) with significantly reduced MOR and KOR affinities, resulting in increased DOR selectivity (MOR/DOR = 306; KOR/DOR = 1009). These trends are consistent with the previous report,²⁹⁾ in which 5′-methoxy substitution preserved DOR affinity, 6′-methoxy significantly reduced DOR affinity, and 7′- and 8′-methoxy led to slight reductions in DOR binding.

Table 1. Binding Affinity of KNT-127 and Its Methoxy-Substituted Derivatives 3a–3d^a)

Compounds	Binding affinity, Ki (nM) (95% CI)a)			Selectivity ratio
	MORb)	DORc)	KORd)	MOR/DOR	KOR/DOR
KNT-127	61.7 (35.6–107)	0.36 (0.26–0.50)	83.2 (46.1–150)	171	231
3a (5′-OMe)	42.0 (14.2–125)	0.12 (0.083–0.16)	15.2 (7.41–31.3)	350	127
3b (6′-OMe)	39.7 (17.5–89.9)	0.79 (0.50–1.23)	54.7 (28.2–106)	50	69
3c (7′-OMe)	113 (45.2–281)	0.64 (0.36–1.14)	290 (186–454)	177	453
3d (8′-OMe)	104 (47.5–229)	0.34 (0.18–0.63)	343 (220–537)	306	1009

a) K_i values were calculated based on receptor-binding-assay results using HEK293 cell membranes. Cell membranes were obtained from HEK293 cells stably expressing each human opioid receptor. Data are presented as the mean (95% confidence interval [95% CI]) calculated from 2 independent experiments, each performed in triplicate. b) [³H]-DAMGO was used. c) [³H]-DPDPE was used. ^d) [³H]-U69,593 was used.

Signal Evaluation Using the NanoBiT Assay

We next assessed the effect of the methoxy substitution of KNT-127 on DOR-mediated Gi-protein dissociation and βarr2 recruitment using the NanoBiT assay^31–33) (Fig. 2 and Table 2). The 5′-methoxy derivative 3a showed approximately 1.7-fold higher Gi-protein dissociation potency compared to KNT-127 (EC_50G = 1.24 nM; E_max = 86.9%), while its βarr2 recruitment potency increased by a factor of 2.4, although the efficacy was halved (3a: EC_50β = 4.66 nM; E_max = 22.5%; KNT-127: EC_50β = 11.2 nM; E_max = 36.7%). The 6′-methoxy derivative 3b maintained the efficacy for Gi-protein dissociation but showed a 294-fold reduction in potency (EC_50G = 620 nM; E_max = 93.0%) as well as substantially decreased βarr2 recruitment potency and efficacy (EC_50β = 1490 nM; E_max = 17.9%). The 7′-methoxy derivative 3c exhibited a 2.9-fold reduced potency for Gi-protein dissociation (EC_50G = 6.14 nM; E_max = 91.0%) but maintained βarr2 recruitment activity similar to that of KNT-127 (EC_50β = 22.3 nM, E_max = 25.9%). The 8′-methoxy derivative 3d showed an efficacy comparable to that of KNT-127 in Gi-protein dissociation, with a 3.5-fold reduction in potency (EC_50G = 7.38 nM; E_max = 86.5%) and a 1.9-fold reduction in βarr2 efficacy (EC_50β = 18.3 nM; E_max = 19.6%).

Fig. 2. Concentration–Response Curves for Gi-Protein Dissociation (Left) and βarr2 Recruitment (Right) Activated through DORs Induced by Met-Enk (Control; Dash), KNT-127 (Black), 3a (Blue), 3b (Red), 3c (Orange), and 3d (Green) at Concentrations of 10^–11–10^–5 M

Table 2. NanoBiT-Based Gi-Protein Dissociation and βarr2 Recruitment Assays for Met-Enk, KNT-127, and Its Methoxy-Substituted Derivatives 3a–3d

Compounds	Gi-protein dissociation		βarr2 recruitment
	EC50 (nM)	Emax (%)a)	EC50 (nM)	Emax (%)b)
Met-Enk	1.31 ± 0.04	100 ± 1.18	34.9 ± 3.18	100 ± 4.51
KNT-127	2.11 ± 0.09	95.6 ± 1.56	11.2 ± 0.14	36.7 ± 1.48
3a (5′-OMe)	1.24 ± 0.14	86.9 ± 0.66	4.66 ± 0.48	22.5 ± 0.52
3b (6′-OMe)	620 ± 39.5	93.0 ± 1.05	1490 ± 100	17.9 ± 1.62
3c (7′-OMe)	6.14 ± 0.41	91.0 ± 0.82	22.3 ± 0.44	25.9 ± 1.80
3d (8′-OMe)	7.38 ± 1.13	86.5 ± 3.18	18.3 ± 0.70	19.6 ± 0.61

a) Gi-protein dissociation E_max was calculated as a relative value toward the reference ligand (Met-Enk) as 100%. b) βarr2 recruitment E_max was calculated as a relative value with Span toward the reference ligand (Met-Enk) as 100%. Data are presented as mean ± S.E.M. from 3 independent experiments, each performed in duplicate.

While the Gi-protein dissociation activity correlated with the DOR binding affinity, βarr2 recruitment was significantly reduced for all derivatives except for the 7′-methoxy derivative 3c. Notably, the 5′-methoxy group enhanced Gi-protein dissociation while reducing βarr2 recruitment efficacy, whereas the 6′-methoxy group weakened both Gi-protein dissociation and βarr2 recruitment.

Docking and Molecular Dynamics Simulations

To gain structural insights into these activity changes, docking studies were conducted using the X-ray crystal structure of the DPI-287–DOR complex²²⁾ (Fig. 3A and Supplementary Fig. S4). Induced-fit docking of KNT-127 produced 17 poses, from among which 6 with docking scores below −8 kcal/mol were selected as templates for grid docking each derivative (Supplementary Fig. S3). Models reflecting the experimental results were selected based on the correlation coefficients between the docking scores and the log K_i values (Supplementary Table S1). In line with our initial prediction, KNT-127 occupied the same binding pocket as NTI, whereby the protonated tertiary 17-amine interacts ionically with D128^3.32, and the quinoline ring occupies the sub-pocket consisting of V281^6.55, W284^6.58, and L300^7.35 (Supplementary Fig. S5). The docking poses of the 5′-, 7′-, and 8′-methoxy derivatives (3a, 3c, and 3d) closely aligned with that of KNT-127, whereas that of the 6′-methoxy derivative 3b intruded further into the pocket, showing an energy increase of >1.3 kcal/mol compared to KNT-127 and the other derivatives. Figure 3B illustrates the distinct docking poses of KNT-127 and 3b, highlighting a root mean square deviation of 1.13 Å between their shared scaffolds. The 6′-methoxy group of 3b causes steric interactions with W284^6.58, tilting the quinoline ring and displacing the N-methylpiperidine moiety deeper into the TM3-TM7 interface.

Fig. 3. (A) Superimposition of the Docking Poses of KNT-127 (Green), 3a (Yellow), 3b (Cyan), 3c (Orange), and 3d (Pink) in the Ligand-Binding Site of DOR

(B) Comparison of the docking poses between KNT-127 (green) and 3b (cyan). (C) Number of frames showing interactions between KNT-127, 3a, and 3b with W284^6.58 and L300^7.35 during the MD simulations.

Molecular dynamics (MD) simulations (1 µs; 200 ns × 5 replicates) for the complex models for KNT-127, 3a, and 3b were conducted to obtain structural insight into the differences in activity of the 5′ and 6′ derivatives. All the ligand–DOR complexes remained stable, supporting the validity of the docking poses (Supplementary Fig. S6). A protein–ligand contact analysis revealed that derivatives 3a and 3b engage in distinct interaction patterns with W284^6.58 and L300^7.35 (Supplementary Fig. S7). The 5′-methoxy derivative 3a and KNT-127 predominantly engage in π–π interactions with W284^6.58, while the 6′-methoxy derivative 3b primarily exhibits hydrophobic interactions, with fewer interaction trajectories involving W284^6.58 (Fig. 3C). The interaction frequency with L300^7.35 was approximately twice as high for 3a as that for KNT-127 (KNT-127: 12156 trajectories; 3a: 31276 trajectories).

These findings suggest that altering the interactions with W284^6.58 through 6′-methoxy substitution in KNT-127 leads to weakened affinity, Gi-protein dissociation, and βarr2 recruitment. In contrast, the increased interaction with L300^7.35 resulting from 5′-methoxy substitution in KNT-127 enhances affinity and Gi-protein dissociation while reducing βarr2 recruitment.

Design, Synthesis, and in Vitro Pharmacology of Novel 5′-Substituted KNT-127 Derivatives 4a and 4b

Molecular Design

Building upon the finding that 5′-methoxy substitution of KNT-127 enhances Gi-protein dissociation while reducing βarr2 recruitment, likely due to interactions with L300^7.35, we hypothesized that introducing bulkier substituents at the 5′-position could further attenuate βarr2 recruitment. To test this hypothesis, we designed and synthesized new derivatives with 5′-isopropyl (iPr; 4a), 5′-ethoxy (OEt; 4b), 5′-isopropoxy (OiPr; 4c), 5′-normalbutoxy (OnBu; 4d), and 5′-cyclohexyloxy (OcHex; 4e) groups and evaluated them by in vitro pharmacology.

Synthesis

To obtain the 5′-iPr derivative 4a, morphinan (5) was treated with 2-amino-6-bromobenzaldehyde under basic conditions to form the quinoline ring in 98% yield, followed by Suzuki–Miyaura coupling to introduce the isopropenyl group, which was reduced to form 8 (81% yield over 2 steps). Finally, demethylation of 8 afforded the 5′-iPr derivative 4a in 67% yield (Chart 1). The 5′-substituted derivatives 4b–4e were synthesized using the method described for 3a–3d above (4b: 71%; 4c: 79%; 4d: 61%; 4e: 52%; Chart 1). Similarly to the methoxy-substituted derivatives 3a–3d, the structures of the synthesized 5′-substituted derivatives 4a–4e were confirmed by 2D-NMR analysis (Supplementary Figs. S12–S16).

Binding Assay

The MOR, DOR, and KOR binding affinities of the synthesized 5′-substituted derivatives 4a–4e were evaluated with competitive binding assays using radiolabeled ligands (Table 3 and Supplementary Fig. S2). The substitution of the 5′-OMe group with a branched iPr group (4a; K_i = 0.56 nM; MOR/DOR = 167; KOR/DOR = 46) resulted in a 4.7-fold reduction in DOR affinity and a slight decrease in receptor selectivity compared to 3a. Additionally, extending the alkyl chain from the 5′-OMe to the 5′-OEt group (4b; K_i = 0.30 nM; MOR/DOR = 36; KOR/DOR = 44) led to a slight reduction in DOR affinity but an increase in MOR affinity, resulting in reduced DOR selectivity. Furthermore, the branched 5′-OiPr derivative (4c; K_i = 0.32 nM; MOR/DOR = 103; KOR/DOR = 57) exhibited comparable DOR affinity to 4b. Further extending the alkyl chain from 5′-OEt to 5′-OnBu (4d; K_i = 0.17 nM; MOR/DOR = 165; KOR/DOR = 95) and from 5′-OiPr to 5′-OcHex (4e; K_i = 0.17 nM; MOR/DOR = 479; KOR/DOR = 202) retained DOR affinity similar to 3a. Notably, 4e demonstrated a decreased affinity for MOR and KOR, resulting in higher DOR selectivity. These results suggest that the introduction of bulky, lipophilic substituents at the 5′-position retains potent DOR affinity and high receptor selectivity.

Table 3. Binding Affinities of 5′-Substituted KNT-127 Derivatives 4a–4e^a)

Compoundsa)	Binding affinity, Ki (nM) (95% CI)a)			Selectivity ratio
	MORb)	DORc)	KORd)	MOR/DOR	KOR/DOR
3a (5′-OMe)	42.0 (14.2–125)	0.12 (0.08–0.16)	15.2 (7.41–31.3)	350	127
4a (5′-iPr)	93.3 (51.3–170)	0.56 (0.39–0.79)	25.6 (18.7–34.9)	167	46
4b (5′-OEt)	10.9 (6.18–19.2)	0.30 (0.23–0.40)	13.1 (9.62–17.9)	36	44
4c (5′-OiPr)	32.9 (17.2–62.9)	0.32 (0.23–0.46)	18.1 (9.77–33.7)	103	57
4d (5′-OnBu)	28.0 (11.6–67.7)	0.17 (0.12–0.24)	16.2 (9.35–28.0)	165	95
4e (5′-OcHex)	81.4 (40.4–164)	0.17 (0.12–0.26)	34.4 (19.2–61.7)	479	202

^a) Ki values were calculated based on receptor-binding-assay results using HEK293 cell membranes. Cell membranes were obtained from HEK293 cells stably expressing each human opioid receptor. Data are presented as the mean (95% confidence interval [95% CI]) calculated from 2 independent experiments, each performed in triplicate. Compounds were evaluated on separate assay plates. KNT-127 was used as a control in each assay to ensure data consistency. The measured K_i values (DOR) for KNT-127 were as follows: 0.36 nM (0.26–0.50 nM) for the plate comparing 3a, 0.29 nM (0.21–0.40 nM) for the plate comparing 4a and 4b, and 0.35 nM (0.27–0.44 nM) for the plate comparing 4c–4e. ^b) [³H]-DAMGO was used. ^c) [³H]-DPDPE was used. ^d) [³H]-U-69593 was used.

Signal Evaluation Using the NanoBiT Assay

Next, the impact of the different 5′-substituents on Gi-protein dissociation and βarr2 recruitment were assessed using the NanoBiT assay (Fig. 4 and Table 4). The 5′-iPr derivative 4a showed a 5.4-fold decrease in potency for Gi-protein dissociation compared to the 5′-OMe derivative 3a while exhibiting a 2.8-fold decrease in βarr2 recruitment potency and a 1.9-fold decrease in efficacy (4a, EC_50G = 6.68 nM; E_max = 70.0%, EC_50β = 13.1 nM; E_max = 11.6%). Extending the 5′-OMe group to the OEt group (4b) resulted in a 2.5-fold reduction in Gi-protein dissociation potency compared to 3a and a 2.2-fold reduction in βarr2 recruitment potency while maintaining similar efficacy (4b, EC_50G = 3.11 nM; E_max = 87.4%, EC_50β = 10.1 nM; E_max = 20.4%). The 5′-OiPr derivative 4c, which contains a branched ether structure, demonstrated a 3.7-fold reduction in Gi-protein dissociation potency compared to 4b, but βarr2 recruitment potency remained similar while efficacy was reduced by 2.0-fold (4c, EC_50G = 4.60 nM; E_max = 76.0%, EC_50β = 8.93 nM; E_max = 11.0%). In contrast, further extending the alkyl group from 5′-OEt to 5′-OnBu decrease in potency for both Gi-protein dissociation and βarr2 recruitment, but enhanced efficacy (4d, EC_50G = 17.7 nM; E_max = 102%; EC_50β = 40.3 nM; E_max = 31.5%). Additionally, the 5′-OcHex derivative 4e exhibited a 14-fold reduction in potency for Gi-protein dissociation compared to the 5′-OiPr derivative 4c, while only the potency for βarr2 recruitment decreased (4e, EC_50G = 64.8 nM; E_max = 98.4%, EC_50β = 39.7 nM; E_max = 13.3%). These results suggest that bulky hydrophobic substituents at the 5′-position of the quinoline ring significantly minimize βarr2 recruitment activity while maintaining potent Gi-protein dissociation activity, probably due to the enhancement of steric interaction with L300^7.35. Notably, branched structures on the ether moiety rather than on the benzylic position reduced βarr2 recruitment activity, whereas further extension of the alkyl group led to decreased Gi-protein dissociation activity. This structural modification approach highlights the potential to fine-tune DOR agonist activity toward desirable therapeutic profiles, paving the way for developing DOR-targeted drugs with reduced side effects.

Fig. 4. Concentration–Response Curves in Gi-Protein Dissociation (Left) and βarr2 Recruitment (Right) through DORs Induced by Met-Enk (Control; Dash), KNT-127 (Gray), 3a (Black), 4a (Blue), 4b (Red) 4c (Orange), 4d (Green), and 4e (Pink) at Concentrations of 10^–11–10^–5 M

Table 4.  NanoBiT-Based Gi-Protein Dissociation and βarr2 Recruitment Assays of 5′-Substituted Derivatives 3a and 4a–4e

Compounds	Gi-protein dissociation		βarr2 recruitment
	EC50 (nM)	Emax (%)a)	EC50 (nM)	Emax (%)b)
Met-Enk	1.31 ± 0.04	100 ± 1.18	34.9 ± 3.18	100 ± 4.51
KNT-127	2.11 ± 0.09	95.6 ± 1.56	11.2 ± 0.14	36.7 ± 1.48
3a (5′-OMe)	1.24 ± 0.14	86.9 ± 0.66	4.66 ± 0.48	22.5 ± 0.52
4a (5′-iPr)	6.68 ± 0.84	70.0 ± 0.23	13.1 ± 0.90	11.6 ± 1.15
4b (5′-OEt)	3.11 ± 0.45	87.4 ± 0.92	10.1 ± 1.20	20.4 ± 1.20
4c (5′-OiPr)	4.60 ± 0.47	76.0 ± 1.39	8.93 ± 1.43	11.0 ± 1.35
4d (5′-OnBu)	17.7 ± 4.63	102 ± 6.53	40.3 ± 9.78	31.5 ± 1.97
4e (5′-OcHex)	64.8 ± 8.74	98.4 ± 0.65	39.7 ± 17.7	13.3 ± 1.87

^a) Gi-protein dissociation E_max was calculated as a relative value toward the reference ligand (Met-Enk) as 100%. ^b) βarr2 recruitment E_max was calculated as a relative value with Span toward the reference ligand (Met-Enk) as 100%. Data are presented as mean ± S.E.M. from 3 independent experiments, each performed in duplicate.

Conclusion

This study has shed light on previously unexplored aspects of the structure–signal relationship in DOR agonists, particularly in terms of minimizing βarr2 recruitment while maintaining Gi-protein activation through targeted structural modifications on the quinoline ring. Through the systematic introduction of a methoxy group or bulkier hydrophobic groups at the 5′-, 6′-, 7′-, and 8′-positions of KNT-127, we determined that the 5′-position plays a critical role in modulating both binding affinity and signaling bias. Specifically, 5′-methoxy substitution improved Gi-protein activity while reducing β-arrestin recruitment; this trend was further enhanced by the introduction of branched structures on the ether moiety, such as isopropoxy (4c) and cyclohexanoxy (4e) at the 5′-position. These modifications effectively preserved DOR binding affinity and selectivity while substantially reducing βarr2 recruitment, suggesting the potential of these derivatives as promising candidates for the development of therapeutics with fewer side effects.

Docking and MD simulations provided insights into the structural interactions responsible for these effects, revealing distinct interaction profiles between the differently substituted quinoline rings and the V281^6.55-W284^6.58-L300^7.35 sub-pocket. The 5′-methoxy and larger hydrophobic 5′-substituents reinforced interactions with L300^7.35, thus stabilizing the ligand orientation in a way that favored G-protein signaling over β-arrestin signaling.

Notably, the morphinan scaffold is recognized as a privileged structure for opioid receptors, where appropriate modifications can achieve receptor subtype selectivity. Beyond the addition of the quinoline ring into the morphinan framework, other beneficial molecular designs for enhancing DOR addressability include the use of oxazatricyclodecane^34,35) and azatricyclodecane³⁶⁾ structures. These scaffolds are known to effectively impart DOR selectivity and potent agonistic activity. However, their effects on signal transduction remain unexplored. The structure–signal relationship framework established in this study provides a valuable approach for optimizing not only quinoline-based morphinan derivatives but also other chemotypes, including oxazatricyclodecane-based morphinan derivatives, to achieve targeted signaling profiles.

While this study has made significant strides toward the rational design of DOR-targeted drugs with optimized efficacy and safety profiles, some limitations remain. The in vitro assays and docking simulations, while valuable for initial insights, do not fully capture the complexity of in vivo environments, in which additional factors, such as receptor desensitization, the cellular context, and pharmacokinetics, could impact the efficacy and side effect profiles. Further in vitro and in vivo studies, as well as extended MD simulations and an analysis of the active-state structure of the complex between these ligands and DOR, will be necessary to validate these findings and establish a comprehensive understanding of the therapeutic potential of these compounds.

In conclusion, this study demonstrates the feasibility of structurally optimizing DOR agonists to achieve desirable signaling bias by modifying the quinoline ring, particularly at the 5′-position. These findings lay the groundwork for the rational design of DOR-targeted therapeutics with enhanced efficacy and safety profiles to address the clinical limitations associated with current opioid analgesics.

Experimental

Chemistry

General Information

All chemical reagents and solvents were purchased from the following commercial suppliers: Tokyo Chemical Industry (Tokyo, Japan), Sigma-Aldrich Inc. (St. Louis, MO, U.S.A.), Kanto Chemical Co., Inc. (Tokyo, Japan), Enamine Ltd. (Monmouth Junction, NJ, U.S.A.), Wako Pure Chemical Corporation Ltd. (Osaka, Japan), and Nacalai Tesque (Kyoto, Japan). All commercially available chemicals and solvents were used without further purification. The synthesis of the compounds described in this study was followed by analytical TLC (Merck Co. Ltd., Darmstadt, Germany; Kieselgel 60 F₂₅₄, 0.25 mm), visualized under 254 nm UV light using phosphomolybdic acid in an aqueous solution of sulfuric acid, Hanessian stain, ninhydrin, or p-anisaldehyde, followed by heating. Column chromatography was carried out on silica gel (a: spherical, neutral, 40–50 µm [Kanto Chemical Co., Japan]; b: spherical, neutral, CHROMATOREX PSQ60B, 60 µm [Fuji Silysia Chemical Ltd., Aichi, Japan]). Preparative TLC was performed on Kieselgel 60 F₂₅₄ plates (Merck Co., Ltd., 0.50 mm). Optical rotations were measured using an Anton Paar MCP 100 Polarimeter. Infrared spectra were recorded using a JASCO (Tokyo, Japan) FT/IR 4100Plus spectrophotometer. ¹H- and ¹³C-NMR spectral data were obtained using JEOL (Tokyo, Japan) JNM-ECS 400 instruments. Chemical shifts are quoted in parts per million (ppm) using CDCl₃ (δ 7.26 ppm) or tetramethylsilane (TMS; δ 0 ppm) as the reference for the ¹H-NMR data and CDCl₃ (δ 77.16 ppm) for the ¹³C-NMR data. Signal patterns are given as br (broad peak), s (singlet), d (doublet), sep (septet), or m (multiplet). ¹H-NMR chemical shifts were assigned using a combination of correlation spectroscopy (COSY), nuclear Overhauser effect spectroscopy (NOESY), and heteronuclear single quantum coherence (HSQC) data. Similarly, ¹³C-NMR chemical shifts were assigned based on HMBC and HSQC experiments. Mass spectra were measured using a JEOL JMS-T100LP (electrospray ionization time-of-flight (ESI-TOF)) mass spectrometer. The purity (≥95%) of the assayed compounds was determined via analytical HPLC or elemental analysis. Analytical HPLC was performed using an ACQUITY ultra-performance liquid chromatography (UPLC) system (Waters Corporation, Milford, MA, U.S.A.) equipped with an ACQUITY UPLC BEH C18 column (1.7 µm, 50 × 2.1 mm); PDA detection at 210–400 nm; column temperature: 40°C.

Synthesis of the Target Compounds and Their Spectroscopic Data

(6R,6aS,14aR)-9-Methoxy-17-methyl-5,6,7,14-tetrahydro-6aH-6,14a-(epiminoethano)naphtho[2,1-b]acridine-2,6a-diol (3a)

A stirred solution of 6 (20 mg, 0.070 mmol) in EtOH (0.7 mL) was treated with methanesulfonic acid (24.6 µL, 0.38 mmol) and 2-amino-6-methoxybenzaldehyde (28.8 mg, 0.190 mmol), before the mixture was stirred at reflux under an Ar atmosphere. After 12 h, the reaction mixture was adjusted to pH = 9 using a saturated aqueous solution of NaHCO₃ (4.0 mL) and extracted with CHCl₃ (20, 10, and 5 mL). The combined organic extracts were dried over Na₂SO₄, filtered, and then the filtrate was concentrated under reduced pressure. The thus obtained residue was purified by column chromatography on silica gel (0–10%, 10% NH₃ aq./MeOH in CHCl₃) and preparative TLC (10%, 10% NH₃ aq./MeOH in CHCl₃) to afford 3a (20.4 mg, 80%) as a colorless amorphous solid (purity ≥95%, determined by UPLC at UV 254 nm).

¹H-NMR (400 MHz, CDCl₃) δ: 8.06 (s, 1H), 7.06 (dd, J = 8.2, 7.8 Hz, 1H), 7.01 (d, J = 2.3 Hz, 1H), 6.97–6.93 (m, 2H), 6.63 (dd, J = 8.2, 2.3 Hz, 1H), 6.52 (d, J = 7.8 Hz, 1H), 3.88 (s, 3H), 3.67 (d, J = 16.9 Hz, 1H), 3.58 (d, J = 16.9 Hz, 1H), 3.26 (d, J = 18.3 Hz, 1H), 3.10 (d, J = 18.3 Hz, 1H), 2.99–2.88 (m, 3H), 2.46–2.42 (m, 4H), 2.31 (ddd, J = 12.8, 11.9, 3.2 Hz, 1H), 2.15 (ddd, J = 12.8, 11.9, 5.5 Hz, 1H), 1.38 (m, 1H), 2 protons (OH) were not observed; ¹³C-NMR (100 MHz, CDCl₃) δ: 157.5, 155.6, 154.4, 146.4, 140.8, 131.0, 128.9, 128.4, 127.5, 126.7, 119.8, 119.3, 115.0, 113.1, 103.4, 69.6, 61.9, 55.7, 45.7, 43.2, 40.6, 38.7, 36.3, 36.2, 24.1. The spectral data were in agreement with those reported in the literature.²⁹⁾

(6R,6aS,14aR)-10-Methoxy-17-methyl-5,6,7,14-tetrahydro-6aH-6,14a-(epiminoethano)naphtho[2,1-b]acridine-2,6a-diol (3b)

A stirred solution of 6 (20 mg, 0.070 mmol) in EtOH (0.7 mL) was treated with methanesulfonic acid (24.6 µL, 0.38 mmol) and 2-amino-5-methoxybenzaldehyde (28.8 mg, 0.190 mmol), before the mixture was stirred at reflux under an Ar atmosphere. After 12 h, the reaction mixture was adjusted to pH = 9 using a saturated aqueous solution of NaHCO₃ (4.0 mL) and extracted with CHCl₃ (20, 10, and 5 mL). The combined organic extracts were dried over Na₂SO₄, filtered, and then the filtrate was concentrated under reduced pressure. The thus obtained residue was purified by column chromatography on silica gel (0–10%, 10% NH₃ aq./MeOH in CHCl₃) and preparative TLC (10%, 10% NH₃ aq./MeOH in CHCl₃) to afford 3b (19.2 mg, 76%) as a colorless amorphous solid (purity ≥95%, determined by UPLC at UV 254 nm).

¹H-NMR (400 MHz, CDCl₃) δ: 7.50 (s, 1H), 7.08 (d, J = 2.3 Hz, 1H), 7.01 (d, J = 8.7 Hz, 1H), 6.78 (d, J = 9.2 Hz, 1H), 6.68–6.65 (m, 2H), 6.57 (dd, J = 9.2, 2.8 Hz, 1H), 3.80 (s, 3H), 3.66–3.62 (m, 2H), 3.31 (d, J = 18.3 Hz, 1H), 3.07 (d, J = 17.9 Hz, 1H), 2.99–2.89 (m, 3H), 2.48–2.44 (m, 4H), 2.36 (ddd, J = 12.8, 11.9, 3.2 Hz, 1H), 2.17 (ddd, J = 12.8, 11.9, 5.5 Hz, 1H), 1.42 (m, 1H), 2 protons (OH) were not observed; ¹³C-NMR (100 MHz, CDCl₃) δ: 157.0, 155.7, 154.6, 141.4, 140.7, 135.0, 129.0, 128.6, 128.1, 127.7, 126.9, 121.5, 115.4, 113.7, 103.8, 69.7, 61.9, 55.3, 45.7, 43.2, 40.6, 38.5, 36.2, 36.0, 24.1. The spectral data were in agreement with those reported in the literature.²⁹⁾

(6R,6aS,14aR)-11-Methoxy-17-methyl-5,6,7,14-tetrahydro-6aH-6,14a-(epiminoethano)naphtho[2,1-b]acridine-2,6a-diol (3c)

A stirred solution of 6 (20 mg, 0.070 mmol) in EtOH (0.7 mL) was treated with methanesulfonic acid (24.6 µL, 0.38 mmol) and 2-amino-4-methoxybenzaldehyde (28.8 mg, 0.190 mmol), before the mixture was stirred at reflux under an Ar atmosphere. After 15 h, the reaction mixture was adjusted to pH = 9 using a saturated aqueous solution of NaHCO₃ (6.0 mL) and extracted with CHCl₃ (20, 10, and 5 mL). The combined organic extracts were dried over Na₂SO₄,afiltered, and then the filtrate was concentrated under reduced pressure. The residue was purified by column chromatography on silica gel (0–10%, 10% NH₃ aq./MeOH in CHCl₃) and preparative TLC (10%, 10% NH₃ aq./MeOH in CHCl₃) to afford 3c (14.3 mg, 56%) as a colorless amorphous solid (purity ≥95%, determined by UPLC at UV 254 nm).

¹H-NMR (400 MHz, CDCl₃) δ: 7.43 (s, 1H), 7.17–7.14 (m, 2H), 7.01 (d, J = 8.2 Hz, 1H), 6.73 (dd, J = 8.2, 2.3 Hz, 1H), 6.65 (dd, J = 9.2, 2.3 Hz, 1H), 6.00 (s, 1H), 3.71–3.62 (m, 2H), 3.33 (d, J = 18.3 Hz, 1H), 3.14 (s, 3H), 3.02–2.88 (m, 4H), 2.51–2.45 (m, 4H), 2.38 (ddd, J = 12.8, 11.9, 3.2 Hz, 1H), 2.21 (ddd, J = 12.8, 11.9, 5.5 Hz, 1H), 1.49 (m, 1H), 2 protons (OH) were not observed; ¹³C-NMR (100 MHz, CDCl₃) δ: 159.6, 157.2, 155.6, 146.6, 140.8, 135.7, 128.9, 127.6, 127.2, 125.6, 122.4, 119.0, 116.3, 115.1, 104.1, 69.6, 61.8, 54.7, 45.6, 43.2, 40.6, 38.6, 36.1, 35.9, 24.1. The spectral data were in agreement with those reported in the literature.²⁹⁾

(6R,6aS,14aR)-12-Methoxy-17-methyl-5,6,7,14-tetrahydro-6aH-6,14a-(epiminoethano)naphtho[2,1-b]acridine-2,6a-diol (3d)

A stirred solution of 6 (20 mg, 0.070 mmol) in EtOH (0.7 mL) was treated with methanesulfonic acid (24.6 µL, 0.38 mmol) and 2-amino-3-methoxybenzaldehyde (28.8 mg, 0.190 mmol), before the mixture was stirred at reflux under an Ar atmosphere. After 12 h, the reaction mixture was adjusted to pH = 9 using a saturated aqueous solution of NaHCO₃ (6.0 mL) and extracted with CHCl₃ (20, 10, and 5 mL). The combined organic extracts were dried over Na₂SO₄, filtered, and then the filtrate was concentrated under reduced pressure. The thus obtained residue was purified by column chromatography on silica gel (1–10%, 10% NH₃ aq./MeOH in CHCl₃) and preparative TLC (10%, 10% NH₃ aq./MeOH in CHCl₃) to afford 3d (23.7 mg, 93%) as a colorless amorphous solid (purity ≥95%, determined by UPLC at UV 254 nm).

¹H-NMR (400 MHz, CDCl₃) δ: 7.61 (s, 1H), 7.15 (dd, J = 8.2, 7.3 Hz, 1H), 7.09 (d, J = 8.2 Hz, 1H), 7.04 (d, J = 2.3 Hz, 1H), 6.95 (d, J = 8.2 Hz, 1H), 6.61 (dd, J = 8.2, 2.3 Hz, 1H), 6.49 (d, J = 7.3 Hz, 1H), 3.77 (d, J = 16.9 Hz, 1H), 3.61 (d, J = 16.9 Hz, 1H), 3.29–3.24 (m, 4H), 3.09–2.88 (m, 4H), 2.48–2.44 (m, 4H), 2.31 (ddd, J = 12.8, 11.9, 3.2 Hz, 1H), 2.18 (ddd, J = 12.8, 11.9, 5.5 Hz, 1H), 1.38 (m, 1H), 2 protons (OH) were not observed; ¹³C-NMR (100 MHz, CDCl₃) δ: 156.6, 155.8, 154.2, 140.8, 137.9, 135.5, 129.0, 128.5, 128.3, 126.7, 125.7, 118.4, 114.9, 113.9, 106.5, 69.6, 61.9, 55.0, 45.8, 43.2, 40.6, 38.9, 36.25, 36.15, 24.1. The spectral data were in agreement with those reported in the literature.²⁹⁾

(6R,6aS,14aR)-9-Bromo-2-methoxy-17-methyl-5,6,7,14-tetrahydro-6aH-6,14a-(epiminoethano)naphtho[2,1-b]acridin-6a-ol (7)

A mixture of 5 (100 mg, 0.317 mmol), 2-amino-6-bromobenzaldehyde (126.9 mg, 0.63 mmol), and 2% KOH/EtOH (0.63 mL) was placed in a sealed microwave vial under an Ar atmosphere. The reaction mixture was heated to 110°C at a pressure of 2 bar under microwave irradiation for 50 min. After cooling to room temperature, the reaction mixture was poured into water (5.0 mL) and extracted with CHCl₃ (20, 10, and 10 mL). The combined organic extracts were dried over Na₂SO₄, filtered, and then the filtrate was concentrated under reduced pressure. The residue was purified by column chromatography on silica gel (0–8%, MeOH in CHCl₃) to afford 7 (145 mg, 98%) as a colorless amorphous solid.

${\left[\alpha \right]}_{\mathrm{D}}^{20}$: −140.3 (c = 0.60, CHCl₃); IR (NaCl): 3426, 2914, 1497, 1273, 1045, 812 cm^–1; ¹H-NMR (400 MHz, CDCl₃) δ: 8.03 (s, 1H), 7.94 (d, J = 8.2 Hz, 1H), 7.66 (d, J = 7.8 Hz, 1H), 7.43 (dd, J = 8.2, 7.8 Hz, 1H), 7.01 (d, J = 8.2 Hz, 1H), 6.88 (d, J = 2.8 Hz, 1H), 6.61 (dd, J = 8.2, 2.8 Hz, 1H), 4.74 (brs, 1H), 3.71 (d, J = 17.4 Hz, 1H), 3.63 (s, 3H), 3.58 (d, J = 17.4 Hz, 1H), 3.27 (d, J = 18.3 Hz, 1H), 3.13 (d, J = 18.3 Hz, 1H), 3.02–2.90 (m, 3H), 2.47–2.41 (m, 4H), 2.29–2.19 (m, 2H), 1.43 (m, 1H); ¹³C-NMR (100 MHz, CDCl₃) δ: 158.42, 158.36, 147.3, 141.0, 134.8, 129.7, 129.3, 128.8, 128.7, 128.4, 127.4, 126.8, 121.1, 112.3, 111.1, 69.5, 61.9, 55.3, 45.8, 43.1, 40.7, 39.3, 36.7, 36.3, 24.2; high-resolution MS (HR-MS) (ESI): m/z [M + H]⁺; Calcd for C₂₅H₂₆BrN₂O₂⁺: 467.11572. Found: 467.11665.

(6R,6aS,14aR)-9-Isopropyl-2-methoxy-17-methyl-5,6,7,14-tetrahydro-6aH-6,14a-(epiminoethano)naphtho[2,1-b]acridin-6a-ol (8)

2-Isopropenyl-4,4,5,5-tetramethyl-1,3,2-dioxaborolane (30.4 µL, 0.16 mmol), Cs₂CO₃ (69.7 mg, 0.21 mmol), and Pd(PPh₃)₄ (24.7 mg, 0.021 mmol) were added to a stirred suspension of 7 (50.0 mg, 0.11 mmol) in 1,4-dioxane (1.0 mL) before the mixture was stirred for 15 h at reflux under an Ar atmosphere. After cooling to room temperature, the reaction mixture was poured into water (5.0 mL) and extracted with CHCl₃ (10, 5, and 5 mL). The combined organic extracts were dried over Na₂SO₄, filtered, and then the filtrate was concentrated under reduced pressure. The thus obtained residue was purified by column chromatography on silica gel (0–8%, MeOH/CHCl₃) to afford the isopropenylated derivative (41.9 mg, 92%) as a yellow amorphous solid.

Then, palladium hydroxide (10.0 mg) was added to a stirred solution of the isopropenylated derivative (10.0 mg, 0.023 mmol) in 1,4-dioxane (1.0 mL) at room temperature under a hydrogen atmosphere. After 15 h of stirring, the reaction mixture was filtered through a pad of Celite, and then the filtrate was concentrated under reduced pressure. The thus obtained residue was purified using preparative TLC (10%, 10% NH₃ aq./MeOH in CHCl₃) to afford 8 (8.8 mg, 88%) as a colorless solid.

${\left[\alpha \right]}_{\mathrm{D}}^{20}$: −150.4 (c = 0.26, CHCl₃); IR (NaCl): 3415, 2927, 1498, 1240, 1043, 753 cm^–1; ¹H-NMR (400 MHz, CDCl₃) δ: 7.94 (s, 1H), 7.83 (d, J = 8.2 Hz, 1H), 7.54 (dd, J = 8.2, 7.3 Hz, 1H), 7.31 (d, J = 7.3 Hz, 1H), 7.00 (d, J = 8.2 Hz, 1H), 6.93 (d, J = 2.8 Hz, 1H), 6.61 (dd, J = 8.2, 2.8 Hz, 1H), 4.73 (brs, 1H), 3.70 (d, J = 17.4 Hz, 1H), 3.63 (s, 3H), 3.58–3.50 (m, 2H), 3.27 (d, J = 17.9 Hz, 1H), 3.11 (d, J = 17.4 Hz, 1H), 2.98–2.90 (m, 3H), 2.45–2.40 (m, 4H), 2.29–2.20 (m, 2H), 1.42 (m, 1H), 1.31 (d, J = 6.9 Hz, 6H); ¹³C-NMR (100 MHz, CDCl₃) δ: 158.4, 156.5, 147.2, 144.2, 141.4, 131.2, 128.6, 128.3, 127.7, 127.5, 126.6, 125.5, 121.2, 112.4, 110.9, 69.7, 61.9, 55.3, 45.7, 43.2, 40.8, 39.5, 36.9, 36.6, 28.3, 24.2, 24.0, 23.5; HR-MS (ESI): m/z [M + H]⁺; Calcd for C₂₈H₃₃N₂O₂⁺: 429.25420. Found: 429.25531.

(6R,6aS,14aR)-9-Isopropyl-17-methyl-5,6,7,14-tetrahydro-6aH-6,14a-(epiminoethano)naphtho[2,1-b]acridine-2,6a-diol (4a)

A stirred solution of 8 (7.7 mg, 0.0179 mmol) in CH₂Cl₂ (0.2 mL) was treated with 1.0 M BBr₃ in CH₂Cl₂ (72.0 µL, 0.36 mmol) at 0°C under an Ar atmosphere. After 1.5 h, the reaction mixture was adjusted to pH = 9 using an aqueous solution of ammonia (30%; 1.0 mL) and extracted with CHCl₃ (20, 10, and 5 mL). The combined organic extracts were dried over Na₂SO₄, filtered, and then the filtrate was concentrated under reduced pressure. The residue was purified using preparative TLC (10%, 10% NH₃ aq./MeOH in CHCl₃) to afford 4b (5.0 mg, 67%) as a colorless solid (purity ≥95%, determined by UPLC at UV max plot).

${\left[\alpha \right]}_{\mathrm{D}}^{20}$: −170.5 (c = 0.20, CHCl₃); IR (KBr): 3401, 2924, 1496, 1284, 755 cm^–1; ¹H-NMR (400 MHz, CDCl₃) δ: 7.94 (s, 1H), 7.29 (dd, J = 7.3, 2.3 Hz, 1H), 7.19–7.14 (m, 2H), 7.01 (d, J = 2.3 Hz, 1H), 6.97 (d, J = 8.2 Hz, 1H), 6.62 (dd, J = 8.2, 2.3 Hz, 1H), 3.69 (d, J = 16.9 Hz, 1H), 3.58 (d, J = 16.9 Hz, 1H), 3.48 (sept, J = 6.9 Hz, 1H), 3.28 (d, J = 18.3 Hz, 1H), 3.13 (d, J = 17.9 Hz, 1H), 2.99–2.89 (m, 3H), 2.45–2.41 (m, 4H), 2.29 (ddd, J = 12.8, 11.9, 3.2 Hz, 1H), 2.17 (ddd, J = 12.8, 11.9, 5.5 Hz, 1H), 1.36 (m, 1H), 1.31 (d, J = 6.9 Hz, 3H), 1.26 (d, J = 6.9 Hz, 3H), 2 protons (OH) were not observed; ¹³C-NMR (100 MHz, CDCl₃) δ: 156.5, 155.3, 146.4, 144.0, 141.1, 131.8, 128.9, 128.4, 127.6, 127.1, 125.4 (2C), 121.3, 114.6, 113.0, 69.7, 61.9, 45.7, 43.2, 40.6, 38.9, 36.5, 36.4, 28.3, 24.2, 24.1, 23.4; HR-MS (ESI): m/z [M + H]⁺; Calcd for C₂₇H₃₁N₂O₂⁺: 415.23855. Found: 415.23961.

(6R,6aS,14aR)-9-Ethoxy-17-methyl-5,6,7,14-tetrahydro-6aH-6,14a-(epiminoethano)naphtho[2,1-b]acridine-2,6a-diol (4b)

A stirred solution of 6 (20 mg, 0.070 mmol) in EtOH (0.7 mL) was treated with methanesulfonic acid (27.1 µL, 0.42 mmol) and 2-amino-6-ethoxybenzaldehyde (34.3 mg, 0.208 mmol), and refluxed under an Ar atmosphere. After 12 h with stirring, the reaction mixture was basified (pH 9) with saturated aqueous NaHCO₃ (6.0 mL), and extracted with CHCl₃ (60, 50, and 40 mL). The combined organic extracts were dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The residue was purified by column chromatography on silica gel (1–10%, 10% NH₃ aq./MeOH in CHCl₃) to afford 4b (20.5 mg, 71%) as a colorless amorphous solid (purity >95%, determined by UPLC at UV 254 nm).

${\left[\alpha \right]}_{\mathrm{D}}^{20}$: −203.2 (c = 0.44, CHCl₃); IR (NaCl): 3392, 2931, 1465, 1264, 755 cm^–1; ¹H-NMR (400 MHz, CDCl₃) δ: 8.09 (s, 1H), 7.03–6.98 (m, 2H), 6.96 (d, J = 8.2 Hz, 1H), 6.85 (d, J = 8.7 Hz, 1H), 6.63 (dd, J = 8.2, 2.3 Hz, 1H), 6.49 (d, J = 7.8 Hz, 1H), 4.14–4.02 (m, 2H), 3.67 (d, J = 16.9 Hz, 1H), 3.61 (d, J = 16.9 Hz, 1H), 3.27 (d, J = 18.3 Hz, 1H), 3.11 (d, J = 17.9 Hz, 1H), 2.99–2.88 (m, 3H), 2.46–2.42 (m, 4H), 2.32 (ddd, J = 12.4, 11.9, 3.2 Hz, 1H), 2.16 (ddd, J = 12.7, 12.4, 5.5 Hz, 1H), 1.47 (dd, J = 6.9, 6.9 Hz, 3H), 1.38 (m, 1H), 2 protons (OH) were not observed; ¹³C-NMR (100 MHz, CDCl₃) δ: 157.5, 155.7, 153.7, 146.4, 140.8, 131.0, 128.9, 128.4, 127.4, 126.8, 119.9, 119.1, 115.0, 113.2, 104.1, 69.6, 63.9, 61.9, 45.7, 43.2, 40.6, 38.7, 36.3, 36.2, 24.1, 14.9; HR-MS (ESI): m/z [M + H]⁺; Calcd for C₂₆H₂₉N₂O₃⁺: 417.21782. Found: 417.21759.

(6R,6aS,14aR)-9-Isopropoxy-17-methyl-5,6,7,14-tetrahydro-6aH-6,14a-(epiminoethano)naphtho[2,1-b]acridine-2,6a-diol (4c)

A stirred solution of 6 (30.0 mg, 0.095 mmol) in EtOH (1.0 mL) was treated with methanesulfonic acid (37.0 µL, 0.57 mmol) and 2-amino-6-isopropoxybenzaldehyde (102.2 mg, 0.57 mmol), before the mixture was stirred at reflux under an Ar atmosphere. After 12 h, the reaction mixture was adjusted to pH = 9 using a saturated aqueous solution of NaHCO₃ (6.0 mL) and extracted with CHCl₃ (20, 10, and 5 mL). The combined organic extracts were dried over Na₂SO₄, filtered, and then the filtrate was concentrated under reduced pressure. The thus obtained residue was purified using column chromatography on silica gel (0–10%, 10% NH₃ aq./MeOH in CHCl₃) and preparative TLC (10%, 10% NH₃ aq./MeOH in CHCl₃) to afford 4a (32.5 mg, 79%) as a colorless amorphous solid (purity ≥95%, determined by UPLC at UV 254 nm).

${\left[\alpha \right]}_{\mathrm{D}}^{20}$: −181.6 (c = 0.55, CHCl₃); IR (KBr): 3401, 2932, 1467, 1263, 1113, 755 cm^–1; ¹H-NMR (400 MHz, CDCl₃) δ: 8.08 (s, 1H), 7.06–7.02 (m, 2H), 6.97 (d, J = 8.2 Hz, 1H), 6.89 (d, J = 8.7 Hz, 1H), 6.62 (dd, J = 8.2, 2.3 Hz, 1H), 6.54 (d, J = 7.3 Hz, 1H), 4.62 (sep, J = 6.0 Hz, 1H), 3.67 (d, J = 16.9 Hz, 1H), 3.60 (d, J = 16.9 Hz, 1H), 3.27 (d, J = 18.3 Hz, 1H), 3.11 (d, J = 18.3 Hz, 1H), 2.98–2.88 (m, 3H), 2.46–2.42 (m, 4H), 2.32 (ddd, J = 12.4, 11.9, 3.2 Hz, 1H), 2.17 (ddd, J = 12.8, 12.4, 5.5 Hz, 1H), 1.40–1.35 (m, 7H), 2 protons (OH) were not observed; ¹³C-NMR (100 MHz, CDCl₃) δ: 157.4, 155.6, 152.6, 146.7, 140.9, 131.2, 128.9, 128.4, 127.3, 126.9, 120.6, 119.0, 114.9, 113.1, 105.5, 70.4, 69.7, 61.9, 45.7, 43.2, 40.6, 38.8, 36.4, 36.2, 24.1, 22.1 (2C); HR-MS (ESI): m/z [M + H]⁺; Calcd. for C₂₇H₃₁N₂O₃⁺: 431.23347. Found: 431.23223.

(6R,6aS,14aR)-9-Butoxy-17-methyl-5,6,7,14-tetrahydro-6aH-6,14a-(epiminoethano)naphtho[2,1-b]acridine-2,6a-diol (4d)

A stirred solution of 6 (20 mg, 0.070 mmol) in EtOH (0.7 mL) was treated with methanesulfonic acid (27.1 µL, 0.42 mmol) and 2-amino-6-butoxybenzaldehyde (40.2 mg, 0.208 mmol), and refluxed under an Ar atmosphere. After 12 h with stirring, the reaction mixture was basified (pH 9) with saturated aqueous NaHCO₃ (6.0 mL), and extracted with CHCl₃ (60, 50, and 40 mL). The combined organic extracts were dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The residue was purified by column chromatography on silica gel (1%–10%, 10% NH₃ aq./MeOH in CHCl₃) to afford 4d (18.9 mg, 61%) as a colorless amorphous solid (purity >95%, determined by UPLC at UV 254 nm).

${\left[\alpha \right]}_{\mathrm{D}}^{20}$: −219.6 (c = 0.49, CHCl₃); IR (NaCl): 3402, 2932, 1465, 1265, 755 cm^–1; ¹H-NMR (400 MHz, CDCl₃): 8.07 (s, 1H), 7.03–6.99 (m, 2H), 6.96 (d, J = 8.2 Hz, 1H), 6.86 (d, J = 8.7 Hz, 1H), 6.63 (dd, J = 8.2, 2.3 Hz, 1H), 6.48 (d, J = 7.8 Hz, 1H), 4.05–3.97 (m, 2H), 3.67 (d, J = 16.9 Hz, 1H), 3.60 (d, J = 16.9 Hz, 1H), 3.27 (d, J = 17.9 Hz, 1H), 3.11 (d, J = 17.9 Hz, 1H), 2.98–2.88 (m, 3H), 2.45–2.41 (m, 4H), 2.31 (ddd, J = 12.4, 11.9, 3.2 Hz, 1H), 2.15 (ddd, J = 12.8, 12.4, 5.5 Hz, 1H), 1.87–1.80 (m, 2H), 1.54 (dq, J = 7.3, 7.3 Hz, 2H), 1.37 (m, 1H), 1.00 (t, J = 7.3 Hz, 3H), 2 protons (OH) were not observed; ¹³C-NMR (100 MHz, CDCl₃) δ: 157.4, 155.7, 153.8, 146.4, 140.8, 131.0, 128.9, 128.5, 127.4, 126.7, 119.9, 119.0, 115.1, 113.2, 104.0, 69.7, 68.0, 61.9, 45.7, 43.2, 40.5, 38.7, 36.3, 36.2, 31.4, 24.1, 19.6, 14.0; HR-MS (ESI): m/z [M + H]⁺; Calcd for C₂₈H₃₃N₂O₃⁺: 445.24912. Found: 445.24966.

(6R,6aS,14aR)-9-(Cyclohexyloxy)-17-methyl-5,6,7,14-tetrahydro-6aH-6,14a-(epiminoethano)naphtho[2,1-b]acridine-2,6a-diol (4e)

A stirred solution of 6 (20 mg, 0.070 mmol) in EtOH (0.7 mL) was treated with methanesulfonic acid (27.1 µL, 0.42 mmol) and 2-amino-6-cyclohexyloxybenzaldehyde (45.7 mg, 0.209 mmol), and refluxed under an Ar atmosphere. After 12 h with stirring, the reaction mixture was basified (pH 9) with saturated aqueous NaHCO₃ (6.0 mL), and extracted with CHCl₃ (60, 50, and 40 mL). The combined organic extracts were dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The residue was purified by column chromatography on silica gel (1–10%, 10% NH₃ aq./MeOH in CHCl₃) to afford 4e (17.0 mg, 52%) as a colorless amorphous solid (purity >95%, determined by UPLC at UV 254 nm).

${\left[\alpha \right]}_{\mathrm{D}}^{20}$: −139.1 (c = 0.32, CHCl₃); IR (NaCl): 3392, 2934, 1467, 1267, 755 cm^–1; ¹H-NMR (400 MHz, CDCl₃) δ: 8.09 (s, 1H), 7.07–7.03 (m, 2H), 6.97 (d, J = 8.2 Hz, 1H), 6.91 (d, J = 8.2 Hz, 1H), 6.63 (dd, J = 8.2, 2.3 Hz, 1H), 6.55 (d, J = 7.8 Hz, 1H), 4.35 (dddd, J = 8.4, 8.4, 4.1, 4.1, 1H), 3.67 (d, J = 16.9 Hz, 1H), 3.59 (d, J = 16.9 Hz, 1H), 3.27 (d, J = 18.3 Hz, 1H), 3.11 (d, J = 17.9 Hz, 1H), 2.98–2.88 (m, 3H), 2.46–2.42 (m, 4H), 2.31 (ddd, J = 12.4, 11.9, 3.2 Hz, 1H), 2.16 (ddd, J = 12.8, 12.4, 5.5 Hz, 1H), 2.03–1.94 (m, 2H), 1.86–1.75 (m, 2H), 1.66–1.54 (m, 3H), 1.46–1.32 (m, 4H), 2 protons (OH) were not observed; ¹³C-NMR (100 MHz, CDCl₃) δ: 157.4, 155.6, 152.5, 146.7, 140.9, 131.2, 128.9, 128.4, 127.3, 126.8, 120.7, 118.9, 114.9, 113.1, 105.7, 75.5, 69.7, 61.9, 45.7, 43.2, 40.6, 38.8, 36.4, 36.2, 31.8, 31.6, 25.8, 24.1, 23.8, 23.7; HR-MS (ESI): m/z [M + H]⁺; Calcd for C₃₀H₃₅N₂O₃⁺: 471.26477. Found: 471.26549.

Bioassays

Competitive Binding Assay

A membrane suspension obtained from HEK293 cells stably expressing the opioid receptors (MOR, DOR, or KOR) was incubated in 250 µL of assay buffer (50 mM Tris, 1 mM ethylenediaminetetraacetic acid (EDTA), 5 mM MgCl₂) with various concentrations of the tested compound and 2 nM tritiated opioid radioligand ([³H]-DAMGO for MOR; [³H]-DPDPE for DOR; [³H]-U69593 for KOR; PerkinElmer Inc., Waltham, MA, U.S.A.) at 25°C for 2 h with gentle shaking at 300 rpm. The reaction was terminated by filtration using a Filtermat B glass filter (PerkinElmer Inc.) with a FilterMate cell harvester (PerkinElmer Inc.). The filters were washed 3 times with 50 mM Tris buffer and then dried for 70 min at 60°C. Finally, MeltiLex B/HS (PerkinElmer Inc.) was melted on the dried filter for 5 min at 90°C. The radioactivity in the filter was determined using a Microbeta scintillation counter (PerkinElmer Inc.). Nonspecific binding was measured in the presence of 10 µM unlabeled opioid ligands (DAMGO for the MORs, DPDPE for the DORs, and U-69593 for the KORs). The sigmoidal concentration–response curves and K_i values were calculated according to the Cheng–Prusoff equation using the Prism software package (Version 8.4.3; GraphPad Software Inc., La Jolla, CA, U.S.A.).

Intracellular Signaling Assays

Reagents and Plasmids

Human full-length DOR was N-terminally fused to the FLAG epitope tag with the preceding hemagglutinin-derived signal sequence (MKTIIALSYIFCLVFADYKDDDDKGGSGGGGSGGSSSGGG; FLAG epitope tag underlined); the resulting construct is referred to as ssHA-FLAG-DOR. Plasmids for the NanoBiT-based G-protein dissociation assay³⁰⁾ and the bystander NanoBiT-based β-arrestin recruitment assay^32,33) have been described previously. Unless otherwise indicated, all the constructs were inserted into the pCAGGS expression plasmid vector.

Cell Culture and Transfection

HEK293A cells (Thermo Fisher Scientific, Waltham, MA, U.S.A.) were maintained in Dulbecco’s modified Eagle’s medium (DMEM; Nissui Pharmaceutical, Tokyo, Japan) supplemented with 5% fetal bovine serum (Gibco, Thermo Fisher Scientific), 100 units/mL penicillin (Sigma-Aldrich), 100 µg/mL streptomycin (Gibco, Thermo Fisher Scientific), and 2 mM l-glutamine (Gibco, Thermo Fisher Scientific) (complete DMEM). Transfection was performed using polyethylenimine (PEI) solution (Polyethylenimine “Max,” Polysciences). Typically, HEK293A cells were seeded in a 10 cm culture dish at a cell density of 2 × 10⁵ cells/mL in 10 mL of the complete DMEM and cultured for 1 d in a humidified incubator (37°C; 5% CO₂). A transfection solution was prepared by combining a plasmid solution diluted in 500 µL of Opti-MEM and 500 µL of Opti-MEM solution containing 25 µg of PEI. The transfected cells were further incubated for 1 d before being subjected to assays as described below.

NanoBiT-Based G-Protein Dissociation Assay

Ligand-induced G-protein dissociation was measured using the NanoBiT-based G-protein dissociation assay,³¹⁾ in which the interaction between a Gα subunit and a Gβγ-subunit was monitored with the NanoBiT system (Promega, Madison, WI, U.S.A.). Specifically, the interaction between the Gαi1 subunit fused with a large fragment (LgBiT) at the α-helical domain (between residues 91 and 92 of Gαi1; Gαi1-LgBiT) and the N-terminally small fragment (SmBiT)-fused Gγ2-subunit with a C68S mutation (SmBiT-Gγ2-CS) was examined. HEK293 cells in a 10 cm dish were transfected with a mixture of 1000 ng of ssHF-DOR plasmid, 500 ng of Gα-LgBiT plasmid, 2500 ng of Gβ1 plasmid, and 2500 ng of SmBiT-Gγ2 (C68S) plasmid. After 24 h of incubation, the transfected cells were collected with 0.53 mM EDTA-containing phosphate-buffered saline, centrifuged at 190g for 5 min and suspended in 2 mL of Hank’s Balanced Salt Solution containing 0.01% bovine serum albumin (BSA; fatty acid-free grade; SERVA) and 5 mM HEPES ((4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid)) (pH = 7.4) (assay buffer). The cell suspension was dispensed in a white 96-well plate at a volume of 80 µL/well and loaded with 20 µL of 50 µM coelenterazine (Carbosynth or Angene) diluted in the assay buffer. After 2 h of incubation at room temperature, the plate was measured for baseline luminescence (SpectraMax L, Molecular Devices), and 20 µL of 6× ligand diluted in the assay buffer, or the assay buffer alone (vehicle), were manually added. The plate was read for 15 min with an integration time of 0.18 s per read and an interval of 20 s at room temperature. The luminescence counts over 10–15 min after ligand addition were averaged and normalized to the initial counts. Concentration–response curves were calculated by dividing the luminescence change at each concentration point by that of the vehicle-only condition.

Bystander NanoBiT-Based βarr2 Recruitment Assay

Ligand-induced β-arrestin recruitment to the plasma membrane was measured using the bystander NanoBiT-based β-arrestin-recruitment assay,^32,33) in which the interaction between β-arrestin and a plasma membrane-anchored split luciferase was monitored using the NanoBiT system (Promega). Specifically, the interaction between the N-terminally small fragment (SmBiT)-fused βarr2 and the plasma membrane-localized large fragment of the split luciferase (LgBiT-CAAX) was examined. A plasmid encoding LgBiT-CAAX was constructed by fusing a human KRAS-derived CAAX motif (GKKKKKKSKTKCVIM) to the C-terminus of LgBiT via a flexible linker (GGSGGGGSGGSSSGG). HEK293 cells in a 10 cm dish were transfected with a mixture of 1000 ng of ssHF-DOR plasmid, 500 ng of SmBiT-β-arrestin, and 2500 ng of LgBiT-CAAX plasmid. The transfected cells were subjected to the same procedure as described in the “NanoBiT-based G-protein dissociation assay” section.

Statistical Analysis

Statistical analysis procedures were performed using the GraphPad Prism 10 software package (GraphPad, San Diego, CA, U.S.A.). For the analysis of the pharmacological parameters, agonist-induced responses were fitted to a 4-parameter sigmoidal concentration–response curve with the Hill slope constrained to an absolute value of <1.5 using the following equation: Y = Bottom + (Top − Bottom)/(1 + 10^((log EC₅₀ − X) × Hill Slope))) (GraphPad Prism 10). The maximum effect (Span) values (“Top” − “Bottom”) and pEC₅₀ values (negative logarithmic values of the half-maximum effective concentration [EC₅₀] values) were calculated from the obtained concentration–response curves. E_max was calculated as a relative value with the Span toward the reference ligand set to 100%.

Computational Studies

Docking Study

An initial model of the wild-type human delta opioid receptor (UniProt: P41143) was generated based on the crystal structure of the active delta opioid receptor complexed with the small molecule agonist DPI-287 (PDB: 6PT3)²²⁾ using Prime (Schrödinger LLC, New York, NY, U.S.A.). This model was further refined for docking simulations with the Protein Preparation Wizard within Maestro (Schrödinger LLC). For KNT-127 and its methoxy derivatives, ionization and energy minimization were conducted using the OPLS3e force field in the LigPrep tool within Maestro (Schrödinger LLC). The minimized structures were then used as input for docking simulations. Docking simulations of KNT-127 into the DOR model were carried out using induced-fit docking with default parameters.³⁷⁾ The grid box for KNT-127 was centered on the center of mass of the reference position of DPI-287. Subsequently, KNT-127 and its methoxy derivatives were redocked into the KNT-127/DOR complex models (up to 20 models) from the induced-fit docking results using Glide XP mode.^38,39) Finally, the optimal poses of KNT-127 and its methoxy derivatives were selected based on structure–activity relationship analysis, evaluating the correlation between docking scores and log K_i values from binding assays.

MD Simulation

Binding models of Rank 4 of the KNT-127 IFD model and Rank 4 docked with methoxy derivatives were subjected to 5 independent MD simulations with different initial velocities for each model using Desmond Version 2.3 (Schrödinger LLC). The OPLS3e force field^40,41) was used for simulations. The initial model structures were placed into a large POPC bilayer following the PPM server criteria⁴²⁾ and SPC water molecules with periodic boundary conditions using an orthorhombic 10 Å layer simulation box. The system was neutralized, and an ionic force of 0.15 M was set by adding Na⁺ and Cl^- ions. After minimization and relaxation of the model, the MD production phase was performed for 200 ns with a time step of 2 fs in the isothermal–isobaric (NPT) ensemble at 300 K and 1 bar using the Langevin thermostat. Long-range electrostatic interactions were computed using the Smooth Particle Mesh Ewald method. MD trajectories were saved every 10 ps for analysis. All system setups were performed using Maestro (Schrödinger LLC). Protein–ligand contact analyses from MD trajectories were also performed using the Simulation Interaction Diagram (Schrödinger LLC).

Acknowledgments

The authors would like to thank Kayo Sato, Shigeko Nakano, and Ayumi Inoue at Tohoku University for their assistance in preparing plasmids and carrying out the NanoBiT-based GPCR signaling assays, as well as Dr. Daisuke Yoshidome and Dr. Tomoko Satoh at Schrödinger Inc. for their assistance and guidance with the use of simulation software.

Funding

This work was supported by JSPS KAKENHI Grants JP21H05111 (to A.I., T.Sa.), JP21H05113 (to A.I.), JP21H05115 (to T.Sa.), JP22K21351 (to T.Sa.), JP23K26793 (to T.Sa.), and JP24K21281 (to A.I.); JST Grants JPMJMI22H5 (to T.Sa. and A.I.), JPMJFR215T (to A.I.), JPMJMS2023 (to A.I.), and SPRING JPMJSP2124 (to K.K. and A.T.); the Japan Foundation for Applied Enzymology (16H007 to T.Sa.); and AMED under Grant Number JP21zf0127005 (to T.Sa.), JP22ama121038 (to A.I.), and JP22zf0127007 (to A.I.). IIIS is supported by the World Premier International Research Center Initiative (WPI), Japan. K.K. and A.T. were supported by JST in the context of the establishment of University Fellowships toward the Creation of Science and Technology Innovation (JPMJFS2106). T.Sa. was supported by the Young Runners in Strategy of Transborder Advanced Researches (TRiSTAR) program from MEXT (Japan).

Author Contributions

Conceptualization: T.Sa. Data curation: T.Sa., T.H., and A.I. Formal analysis: K.K., T.Su., R.Ki., R.S., and Y.N. Funding acquisition: T.Sa. and A.I. Synthesis: K.K., T.Su., Y.S., and T.K. Binding assays: K.K., R.Ka., and A.T. NanoBiT assays: R.S. and R.Ki. Simulation: T.K., T.Sa., and T.H. Methodology: T.Sa., T.H., and A.I. Project administration: T.Sa. Resources: T.Sa., T.H., and A.I. Supervision: T.Sa. Validation: Y.N. and N.K. Visualization: K.K, T.Su, T.K., R.S., and R.Ki. Original draft preparation: K.K., T.Su. T.K., T.H., and T.Sa. Review and editing: K.K., T.Su., R.Ki., A.I., T.H., and T.Sa. All authors have read and agreed to the published version of the manuscript.

Conflict of Interest

The authors declare no conflict of interest.

Data Availability

Data are presented within the manuscript. Additional raw data are available on request from the corresponding author. Samples of the compounds are available upon request from the authors.

Supplementary Materials

This article contains supplementary materials.

References

• 1)  Martin W. R., Eades C. G., Thompson J. A., Huppler R. E., Gilbert P. E., J. Pharmacol. Exp. Ther., 197, 517–532 (1976).
• 2)  Epstein D. H., Heilig M., Shaham Y., Trends Pharmacol. Sci., 39, 911–916 (2018).
• 3)  Gallantine E. L., Meert T. F., Basic Clin. Pharmacol. Toxicol., 97, 39–51 (2005).
• 4)  Abdallah K., Gendron L., Handb. Exp. Pharmacol., 247, 147–177 (2018).
• 5)  Gavériaux-Ruff C., Kieffer B. L., Behav. Pharmacol., 22 (5 and 6), 405–414 (2011).
• 6)  Quirion B., Bergeron F., Blais V., Gendron L., Front. Mol. Neurosci., 13, 52 (2020).
• 7)  Bertels Z., Witkowski W. D., Asif S., Siegersma K., van Rijn R. M., Pradhan A. A., Headache, 61, 170–178 (2021).
• 8)  Erbs E., Faget L., Scherrer G., Matifas A., Filliol D., Vonesch J. -L., Koch M., Kessler P., Hentsch D., Birling M. -C., Koutsourakis M., Vasseur L., Veinante P., Kieffer B. L., Massotte D., Brain Struct. Funct., 220, 677–702 (2015).
• 9)  Filliol D., Ghozland S., Chluba J., Martin M., Matthes H. W. D., Simonin F., Befort K., Gavériaux-Ruff C., Dierich A., LeMeur M., Valverde O., Maldonado R., Kieffer B. L., Nat. Genet., 25, 195–200 (2000).
• 10)  Saitoh A., Kimura Y., Suzuki T., Kawai K., Nagase H., Kamei J., J. Pharmacol. Sci., 95, 374–380 (2004).
• 11)  Hudzik T. J., Howell A., Payza K., Cross A. J., Eur. J. Pharmacol., 396, 101–107 (2000).
• 12)  Mabrouk O. S., Handb. Exp. Pharmacol., 247, 261–275 (2018).
• 13)  Saitoh A., Nagase H., Handb. Exp. Pharmacol., 247, 3–19 (2016).
• 14)  Richards E. M., Mathews D. C., Luckenbaugh D. A., Ionescu D. F., Machado-Vieira R., Niciu M. J., Duncan W. C., Nolan N. M., Franco-Chaves J. A., Hudzik T., Maciag C., Li S., Cross A., Smith M. A., Zarate C. A. Jr., Psychopharmacology (Berl.), 233, 1119–1130 (2016).
• 15)  Bilsky E. J., Calderon S. N., Wang T., Bernstein R. N., Davis P., Hruby V. J., McNutt R. W., Rothman R. B., Rice K. C., Porreca F., J. Pharmacol. Exp. Ther., 273, 359–366 (1995).
• 16)  Blaine A. T., van Rijn R. M., Neuropharmacology, 232, 109526 (2023).
• 17)  Berthiaume S., Abdallah K., Blais V., Gendron L., J. Neural Transm. (Vienna), 127, 661–672 (2020).
• 18)  Pradhan A. A., Befort K., Nozaki C., Gavériaux-Ruff C., Kieffer B. L., Trends Pharmacol. Sci., 32, 581–590 (2011).
• 19)  Nozaki C., Nagase H., Nemoto T., Matifas A., Kieffer B. L., Gaveriaux-Ruff C., Br. J. Pharmacol., 171, 5376–5386 (2014).
• 20)  Pradhan A. A., Perroy J., Walwyn W. M., Smith M. L., Vicente-Sanchez A., Segura L., Bana A., Kieffer B. L., Evans C. J., J. Neurosci., 36, 3541–3551 (2016).
• 21)  Chiang T., Sansuk K., van Rijn R. M., Br. J. Pharmacol., 173, 332–343 (2016).
• 22)  Claff T., Yu J., Blais V., Patel N., Martin C., Wu L., Han G. W., Holleran B. J., Van der Poorten O., White K. L., Hanson M. A., Sarret P., Gendron L., Cherezov V., Katritch V., Ballet S., Liu Z. -J., Müller C. E., Stevens R. C., Sci. Adv., 5, eaax9115 (2019).
• 23)  Conibear A. E., Asghar J., Hill R., Henderson G., Borbely E., Tekus V., Helyes Z., Palandri J., Bailey C., Starke I., von Mentzer B., Kendall D., Kelly E., J. Pharmacol. Exp. Ther., 372, 224–236 (2020).
• 24)  Meqbil Y. J., Aguilar J., Blaine A. T., Chen L., Cassell R. J., Pradhan A. A., van Rijn R. M., J. Pharmacol. Exp. Ther., 389, 301–309 (2024).
• 25)  Meqbil Y. J., Su H., Cassell R. J., Mores K. L., Gutridge A. M., Cummins B. R., Chen L., van Rijn R. M., Molecules, 26, 7236 (2021).
• 26)  Cheng L., Miao Z., Liu S., Li Z., Fu H., Xu C., Hu S., Zhao C., Liu Y., Zhao T., Liu W., Wang H., Liu R., Yan W., Tang X., Liu J., Shao Z., Ke B., Nat. Commun., 15, 8284 (2024).
• 27)  Kajino K., Sugai T., Kise R., Suzuki R., Tokuda A., Sekiya Y., Kakumoto T., Katamoto R., Kutsumura N., Nagumo Y., Inoue A., Saitoh T., Structure-signal relationships of the δ-opioid-receptor (DOR)-selective agonist KNT-127—Part I: Impact of the morphinan skeleton on the G-protein-biased DOR agonism. ChemRxiv 2024.
• 28)  Granier S., Manglik A., Kruse A. C., Kobilka T. S., Thian F. S., Weis W. I., Kobilka B. K., Nature (London), 485, 400–404 (2012).
• 29)  Ida Y., Matsubara A., Nemoto T., Saito M., Hirayama S., Fujii H., Nagase H., Bioorg. Med. Chem., 20, 5810–5831 (2012).
• 30)  Nagase H., Nemoto T., Matsubara A., Saito M., Yamamoto N., Osa Y., Hirayama S., Nakajima M., Nakao K., Mochizuki H., Fujii H., Bioorg. Med. Chem. Lett., 20, 6302–6305 (2010).
• 31)  Inoue A., Raimondi F., Kadji F. M. N., Singh G., Kishi T., Uwamizu A., Ono Y., Shinjo Y., Ishida S., Arang N., Kawakami K., Gutkind J. S., Aoki J., Russell R. B., Cell, 177, 1933–1947.e25 (2019).
• 32)  Janetzko J., Kise R., Barsi-Rhyne B., Siepe D. H., Heydenreich F. M., Kawakami K., Masureel M., Maeda S., Garcia K. C., von Zastrow M., Inoue A., Kobilka B. K., Cell, 185, 4560–4573.e19 (2022).
• 33)  Xu Z., Ikuta T., Kawakami K., Kise R., Qian Y., Xia R., Sun M. -X., Zhang A., Guo C., Cai X. -H., Huang Z., Inoue A., He Y., Nat. Chem. Biol., 18, 281–288 (2022).
• 34)  Fujii H., Hayashida K., Saitoh A., Yokoyama A., Hirayama S., Iwai T., Nakata E., Nemoto T., Sudo Y., Uezono Y., Yamada M., Nagase H., ACS Med. Chem. Lett., 5, 368–372 (2014).
• 35)  Hayashida K., Hirayama S., Iwai T., Watanabe Y., Takahashi T., Sakai J., Nakata E., Yamakawa T., Fujii H., Nagase H., Bioorg. Med. Chem. Lett., 27, 2742–2745 (2017).
• 36)  Watanabe Y., Hayashida K., Saito D., Takahashi T., Sakai J., Nakata E., Kanda T., Iwai T., Hirayama S., Fujii F., Yamakawa T., Nagase H., Bioorg. Med. Chem. Lett., 27, 3495–3498 (2017).
• 37)  Sherman W., Day T., Jacobson M. P., Friesner R. A., Farid R., J. Med. Chem., 49, 534–553 (2006).
• 38)  Friesner R. A., Banks J. L., Murphy R. B., Halgren T. A., Klicic J. J., Mainz D. T., Repasky M. P., Knoll E. H., Shelley M., Perry J. K., Shaw D. E., Francis P., Shenkin P. S., J. Med. Chem., 47, 1739–1749 (2004).
• 39)  Halgren T. A., Murphy R. B., Friesner R. A., Beard H. S., Frye L. L., Pollard W. T., Banks J. L., J. Med. Chem., 47, 1750–1759 (2004).
• 40)  Roos K., Wu C., Damm W., Reboul M., Stevenson J. M., Lu C., Dahlgren M. K., Mondal S., Chen W., Wang L., Abel R., Friesner R. A., Harder E. D., J. Chem. Theory Comput., 15, 1863–1874 (2019).
• 41)  Kurki M., Poso A., Bartos P., Miettinen M. S., J. Chem. Inf. Model., 62, 6462–6474 (2022).
• 42)  Lomize A. L., Todd S. C., Pogozheva I. D., Protein Sci., 31, 209–220 (2022).