Michael Acceptor Approach to the Design of New Salvinorin A-based High Affinity Ligands for the Kappa-Opioid Receptor

Abstract

The neoclerodane diterpenoid salvinorin A is a major secondary metabolite isolated from the psychoactive plant Salvia divinorum. Salvinorin A has been shown to have high affinity and selectivity for the κ-opioid receptor (KOR). To study the ligand–receptor interactions that occur between salvinorin A and the KOR, a new series of salvinorin A derivatives bearing potentially reactive Michael acceptor functional groups at C-2 was synthesized and used to probe the salvinorin A binding site. The κ-, δ-, and μ-opioid receptor (KOR, DOR and MOR, respectively) binding affinities and KOR efficacies were measured for the new compounds. Although none showed wash-resistant irreversible binding, most of them showed high affinity for the KOR, and some exhibited dual affinity to KOR and MOR. Molecular modeling techniques based on the recently-determined crystal structure of the KOR combined with results from mutagenesis studies, competitive binding, functional assays and structure–activity relationships, and previous salvinorin A–KOR interaction models were used to identify putative interaction modes of the new compounds with the KOR and MOR.

1. Introduction

The neoclerodane diterpenoid salvinorin A (1; Figure 1) is a major metabolite isolated from the leaves of Salvia divinorum [1]. It is a highly selective κ-opioid receptor (KOR) agonist and is the most potent naturally-occurring hallucinogen [2]. It gained significant scientific interest as the only non-nitrogenous KOR agonist with no apparent structural similarity to other ligands. This has encouraged several research groups to study the structure–activity relationships of 1, and a plethora of salvinorin A derivatives has been synthesized over the past decade [3-8]. Some of these analogues exhibit interesting pharmacological profiles from full KOR agonist to partial DOR or MOR agonist and antagonists. However, most of these derivatives displayed decreased affinity or even no affinity for the KOR. Our current objective was to use the knowledge about salvinorin A–KOR interactions to rationally design salvinorin A derivatives with different pharmacological profiles and therapeutic potential. In the course of our work on the molecular mechanism of interaction of salvinorin A with the KOR, we reported the irreversible binding of 22-thiocyanatosalvinorin A (2) (Fig. 1) with the sulfhydryl group of C315 at the κ-opioid receptor [9, 10]. Previous results using the KOR model and crystal structure to predict the mode of binding of 2, combined with mass spectrometric analysis of the enzymatically digested KOR–2 complex show that Cys315_7.35 in the orthosteric binding site is an anchoring amino acid for compounds bearing a thioltrapping group [9, 10]. We therefore designed a series of salvinorin B-based (3) α,β-unsaturated esters at the C-2 position with diverse stereoelectronic properties and whose reactive nucleophilic center positions approximately matched those of the previouslyreported thiol-trapping analogs [10] (Schemes 1 and 2).

Figure 1.

The structures of salvinorin A (1), 22-thiocyanatosalvinorin A (RB-64; 2) and salvinorin B (3).

Scheme 1.

Reagents and Conditions: (a) acryloyl chloride or methacryloyl chloride, Et₃N, dry DCM, 0 °C → rt, 3 h. (b) 3,3-dimethylacryloyl chloride or crotonyl chloride or 2,3-dimethylacryloyl chloride, Et₃N, dry DCM, 0 °C → rt, 3 h.

Scheme 2.

Reagents and Conditions: (a) appropriate cinnamic acid or heterocyclic acrylic acid, DCC, DMAP, dry DCM, rt, 2–5 h.

In this work we describe the synthesis of various potential Michael acceptor-type derivatives of salvinorin A and their affinity to KOR, as well as molecular models of their interaction with the receptor. Michael acceptors are capable of forming irreversible bonds with biological macromolecules such as proteins or DNA [11-15]. Thus, the ability of salvinorin A-derived Michael acceptors to form covalent bonds may be advantageous for probing the affinity to the KOR. Considering the fact that Michael acceptors may also react reversibly with nucleophiles [15-19], such compounds may provide potential drug candidates for CNS disorders. Herein, we report the synthesis of a new series of potential Michael acceptor-type salvinorin A derivatives modified at C-2, a position that has been found to be crucial for a high-affinity binding to the KOR [3]. Among the compounds synthesized and tested were a set of cinnamic acid derivatives, which were selected with the knowledge that salvinorin derivatives bearing an aromatic substituent at the C-2 position (the most noteworthy of which is 2-O-benzoylsalvinorin B or herkinorin[20]) exhibit a marked change in the compounds’ pharmacological profile compared to salvinorin A; in particular, these compounds tended to exhibit reduced affinity for KOR and increased affinity for MOR [3]. At the same time we hypothesized that adding a conjugated bond to the aromatic system might further increase MOR affinity as the introduction of alkene moieties at the C-2 position showed a similar shift in pharmacological profile [21].

2. Results and discussion

2.1.1. Chemistry

The starting material, salvinorin A (1), was isolated from commercially available dried leaves of Salvia divinorum, and then converted to salvinorin B (3) in high yield [22]. The reaction of 3 with appropriate acryloyl chlorides under basic conditions afforded 4a–4h, and with various substituted cinnamic acids the corresponding 5a–5n derivatives as shown in Schemes 1 and 2.

The reaction of 3 with acryloyl chloride or methacryloyl chloride in dry DCM in the presence of triethylamine afforded 4a or 4b [21] in 73 and 70% yield, respectively. The ¹H NMR data of acrylate 4a displayed typical acrylate protons as a doublet of doublets at δ 6.20 (J = 8.0, 16.4 Hz), and two doublets at δ 6.49 (J = 16.4 Hz), and 5.93 (J = 8.0 Hz). The corresponding data of 4b [21] confirmed the presence of the methacrylate group [δ 6.23 (1H, d, J = 1.6 Hz), 5.68 (1H, d, J = 1.6 Hz), and 1.99 (3H, s)]. Interestingly, when 3 was subjected to 3,3-dimethylacryloyl chloride (senecioyl chloride) under the same conditions it yielded two isomers 4c and 4d (Scheme 1). Compound 4c, the anticipated salvinorin 3,3-dimethylacrylate displayed characteristic ¹H- and ¹³C NMR signals at δ 5.80 [1H, s, (CH₃)₂C=CH-CO₂-], 2.17 and 1.92 [each 3H, s, (CH₃)₂C=CH-CO₂-], δ 114.85 (C-2′), 27.35 and 20.46 (each CH₃). The second product (4d) proved to be the non-conjugated isomer of 4c. Its ¹H NMR spectrum displayed the side chain protons at δ 3.15 (H-2′, s), δ 1.85 (CH₃, s), and the olefinic methylene protons at δ 4.90 and 4.95 (both s). ¹³C NMR experiments confirmed the C-2′ and C-4′ methylene carbons at δ 43.20 and 115.17, and the C-3′ methyl carbon at δ 22.46. The formation of conjugated and non-conjugated isomers in esterification of alcohols with 3,3- dimethylacryloyl chloride in the presence of triethylamine was observed earlier [23, 24]. The elimination-addition and competitive substitution mechanisms are responsible for the formation of both isomers. In the first case the reaction proceeds through the formation of a ketene intermediate, in the second case through the acyl quaternary ammonium intermediate. Participation of the elimination-addition mechanism was further supported by using an excess of triethylamine. Changing the excess of Et₃N from 1.5 to 3.0 eq altered the ratio of conjugated (4c) to non-conjugated (4d) product from 1:4 to 1:1. Similarly, a mixture of conjugated and non-conjugated isomers 4e–4h were obtained in the reaction of 3 with crotonyl- and 2,3-dimethylacryloyl chloride (Scheme 1). Salvinorin B (3) treated with aromatic and heterocyclic α, β-unsaturated carboxylic acids in the presence of DCC and DMAP yielded corresponding products 5a–5n in good yields (61– 92%), e.g., reaction of 3 with trans-cinnamic acid at room temperature afforded 2-O-cinnamoylsalvinorin B (5a) in 92% yield (Scheme 2). The ¹H- and ¹³C NMR data of 5a showed the trans olefinic protons at δ 6.52 and 7.75 (each d, J = 16.8 Hz), and the C-2′ and C-3′ olefinic carbons at δ 139.46 and 143.64, respectively.

To enhance our understanding of the binding mode between synthesized ligands and the receptor (reversible vs. irreversible labeling), we performed the model Michael addition reactions between acrylate derivative (4a) and thiols like cysteamine and N-acetylcysteine using the reported procedure [16]. Thus, 4a was dissolved in DMSO-d₆, the ¹H NMR spectrum was recorded (Figure 2b), and then two equivalents of cysteamine were added and the spectrum recorded after 5 min (Figure 2c). This spectrum showed the disappearance of the olefinic protons of the acrylate moiety. An aliquot of the solution was transferred to another vial, and diluted with CDCl₃ (1:15 v/v). The ¹H NMR spectrum (Figure 2d) of this solution showed that the proton resonances of the olefinic system did not reappear, hence indicating the irreversible formation of the covalent thioether bond. Similar results were observed in the experiment with N-acetylcysteine.

Figure 2.

a) The ¹H NMR spectrum of 4a in CDCl₃. b) The ¹H NMR spectrum of 4a in DMSO-d₆. c) The ¹H NMR spectrum of compound 4a in DMSO-d₆ after the addition of cysteamine (note disappearance of the olefin proton resonances). d) The ¹H NMR spectrum of the reaction mixture (acrylate derivative + cysteamine + DMSO-d₆) after dilution with CDCl₃ (sample locked in DMSO-d₆). e) ¹H NMR spectrum of the reaction mixture after dilution with CDCl₃ (sample locked in CDCl₃).

2.1.2 Biological affinity and efficacy at opioid receptors

The synthesized analogues (4a–4h and 5a–5n) were evaluated for κ-, δ-, μ- opioid receptor binding affinities at the NIMH-sponsored Psychoactive Drug Screening Program, University of North Carolina at Chapel Hill. The assays were conducted according to the procedure described earlier [2, 25] and the results are summarized in Table 1. Interestingly, none of the compounds exhibited wash-resistant irreversible binding at KOR as was previously observed for the 22-thiocyanatosalvinorin A analog RB-64 [10]. There may be several reasons for this. First, although the acrylate 4a reacted readily with small thiols in solution (vide supra), the relative position and orientation of the electrophilic center in the acceptor moiety and the C315 sulfur nucleophile in the bound receptor–ligand complex may preclude nucleophilic attack. The relatively rigid and planar character of the acrylate moiety differs from that of the more flexible 22- thiocyanato moiety of RB-64 in which there are two rotatable bonds between the carbonyl carbon atom and the thiocyanate sulfur atom. Second, many of the acrylate derivatives are substituted at the electrophilic carbon; these substituents may bind in separate subpockets of the receptor (vide infra), which may also serve to orient the acceptor moiety away from the nucleophilic sulfur atom. Third, the alkene moiety of the non-conjugated products (4d, 4f, 4h) may not be sufficiently reactive toward sulfur nucleophiles to form a covalent linkage. Fourth, the reversibility of the addition reaction may be too great to detect transient covalent linkages between receptor and ligand. Despite their non-reactivity, however, many of the compounds showed high affinity for KOR. The aliphatic conjugated and non-conjugated series of analogs (4a–h) are all of similar size but with varying patterns of methyl group substitution and unsaturation; each showed significant affinity for the κ-opioid receptor. In the conjugated series, the acrylate derivative 4a had a 6-fold reduced but still appreciable affinity at the κ- receptor as compared to 1 (K_i = 18.1 nM vs. K_i = 2.9 nM) (Table 1). Previously published methacrylate derivative 4b also showed good affinity for the κ- receptor [21]. Owing to the presence of bulky methyl groups at the α- and β-positions, the KOR affinities of 4c and 4g were reduced 21- and 8-fold, respectively, as compared to 4a. The crotonate derivative 4e, however, exhibited almost equal affinity as 4a (K_i = 25.3 nM vs. K_i = 18.1 nM). Interestingly, non-conjugated derivatives 4d and 4f showed 3- and 2-fold increased affinities at the KOR, compared to acrylate derivative 4a (K_i = 6 and 10 nM, respectively, vs. K_i = 18.1 nM). Again, derivative 4h having a methyl group α to the carbonyl had reduced affinity at the KOR as compared to 4a (K_i = 28.8 nM vs. K_i = 18.1 nM). The results show that the non-conjugated series of analogues have higher binding KOR affinities than the corresponding conjugated series of compounds, and that C-α substitution results in lower affinity. This may emphasize the importance of the presence of an intervening methylene group that effectively separates the π-systems of the ester and alkene portions of the side chain over direct ester–alkene conjugation on the affinity to the KOR. Of note is the retention of high KOR selectivity of compounds 4a–4h with low or no affinity to the DOR and MOR, as observed for the smaller analog, salvinorin A [2].

Table 1.

Binding affinities of potential Michael acceptor-type salvinorin A derivatives 4a–4h and 5a–5n at MOR, DOR and KOR and adenylate cyclase inhibition in live HEK293 cells using a cAMP biosensor.

		Affinityc				Inhibitiong
compound		Ki ± SD [nM]		Selectivity		EC50±SD [nM]
	MORd	DORe	KORf	MOR/KOR	DOR/KOR	KOR
DAMGO	2.3±1.1	NDa	NDa	—	—	NDa
naltrindole	NDa	0.9±0.3	NDa	—	—	NDa
1	NDa	NDa	2.9±1.2	—	—	0.36±0.04
4a	979±35	>10000	18.1±5.0	54	>552	137±26
4b	>10000	7823±105	34b±10	>293	221	NDa
4c	284±13	2625±100	385±37	0.73	6.8	698±138
4d	362±35	>10000	6±2	60	>1666	25±5
4e	540±43	>10000	25.3±6.0	21	>395	11±2
4f	1441±104	3260±52	10±4	144	326	23±4
4g	>10000	>10000	153±18	>65	>65	477±81
4h	439±28	3032±135	28.8±8.0	15	105	37±6
5a	52±9	700.5±101	9.6±2.0	5.4	73	7±1
5b	>10000	>10000	64.9±5.0	>154	>154	10±1
5c	>10000	>10000	134.8±17.0	>74	>74	21±2
5d	>10000	>10000	581±13	>17	>17	38±6
5e	>10000	>10000	161.5±12.0	>62	>62	27±3
5f	>10000	>10000	>10000	1	1	1799±352
5g	2820±113	>10000	323±21	8.7	>31	50±5
5h	>10000	>10000	148.3±7.0	>67	>67	33±4
5i	>10000	>10000	280.5±24.0	>35	>35	32±4
5j	>10000	>10000	129.5±9.0	>77	>77	20±3
5k	>10000	>10000	>10000	1	1	642±120
5l	>10000	5871±113	>10000	1	<0.58	1210±223
5m	334±18	4888±100	>10000	<0.03	<0.48	1762±344
5n	18±5	228.2±9.0	304±11	0.06	0.75	582±111

^aNot determined,

^bRef. [21],

^cData are mean of three experiments performed in duplicate,

^dThe binding affinity constant (K_i) determined against [³H]DAMGO ligand,

^eK_i determined against [³H]DADLE ligand.

^fK_i determined against [³H]U69,593 ligand,

^gData are mean of three experiments.

The aromatic series of compounds 5a–5n generally displayed lower affinity for the KOR than the smaller conjugated compounds 4a–c, 4e and 4g. Interestingly, the unsubstituted cinnamate derivative 5a displayed dual affinity at the KOR and MOR, exhibiting 2- and 18-fold increased affinity, respectively, compared to 4a. Compounds with electron withdrawing or donating substituents at different positions of the phenyl ring (5b–5l) showed reduced or no affinity at the KOR and no appreciable affinity for MOR or DOR. For the KOR, several trends could be identified. Substitution at the ortho position is tolerated more than substitution at the para position (compare 5b with 5c, 5h with 5i, and 5j with 5k). The 2,5-dimethoxy analog 5d exhibited a much lower affinity than the 2-methoxy analog 5b, indicating that substitution at the meta position also adversely affects binding affinity. The electron-donating methoxy group is tolerated more than the electron-withdrawing nitro or trifluoromethyl groups (compare 5b with 5h and 5j; also 5c with 5i and 5k). The 3,4,5-trimethoxy compound 5f showed no appreciable affinity for any of the three opioid receptors. Introduction of heteroaromatic ring moieties such as 3-pyridyl (5m) and 3-thiophenyl (5n) groups resulted in a significant increase of the MOR and DOR affinities in comparison to all analogues in the substituted aromatic series, including the phenyl analog 5a. Intriguingly, substitution of the phenyl group with the isosteric 3-pyridyl group completely reverses the KOR–MOR selectivity profile, with analog 5m showing no affinity for KOR but with significant affinity for MOR (K_i = 334 nM). The isosteric thiophenyl derivative 5n also showed high MOR selectivity (K_i = 18 nM) but also had affinity for KOR (K_i = 304 nM) and DOR (K_i = 228 nM).

The efficacy of salvinorin A (1) and analogs 4a–h and 5a–n at the KOR was also evaluated using an assay that measures cAMP inhibition (Table 1). None of the analogs 4a–h or 5a–n were as efficacious as salvinorin A (1; IC₅₀ = 0.36 nM). Many of the tested analogs showed an inhibition level that was ~100× less than that produced by 1, the most efficacious of which (5a; IC₅₀ = 7 nM) showed an inhibition level that was 20× lower than 1. The compounds showing the poorest efficacy (5f, 5l, 5m) also had no substantial affinity for the KOR. Although there was not a strong correlation between affinity and efficacy for the tested compounds (see Figure S1), those of the aromatic series (except 5a) tended to show efficacy similar to those of the non-conjugated series but with lower binding affinity. Interestingly, compound 4a, whose binding affinity is only 6× less than 1, was 380× less efficacious.

2.1.3. Molecular Modeling

We have used the crystal structure of the KOR complexed with the antagonist JDTic (PDB ID = 4DJH) to model the interaction of various agonists with the receptor. Although technically in an inactive state, it has been shown that large-scale movements of amino acid residues in the orthosteric binding site of class A GPCRs are not requisite for activation of the receptor [26]. In addition, it has been found that virtual screening for ligands using inactive-state GPCRs can produces agonist leads (see for example Negri et al.) [27]. We have previously described a putative binding mode for salvinorin A representing an initial recognition mode based on an extensive amount of experimental data [9, 28]. This binding mode was reproduced in the current docking studies and is shown in Figure 3a. In the proposed binding mode, the fused tricyclic core is oriented toward TM2 and interacts primarily with the side chains of V108^2.53, T111^2.56, Q115^2.60, V118^2.63, 134^3.28, 135^3.29 and D138^3.32, which form a highly complementary stereoelectronic binding site for the salvinorin core structure (numbers in parentheses indicate Ballesteros–Weinstein indexes [29]). The furanyl moiety interacts with V118^2.63, W124^EL1 and C210^EL2, and in a slightly different pose, may also possibly form hydrogen bonds with Y313^7.36.9 The C4 methyl ester is situated in a small pocket formed by V108^2.53, T111^2.56, W287^6.48 and Y320^7.43 and engages in hydrogen bonding interactions with T111^2.56 and Y320^7.43. The C2 acetoxy group engages a small subpocket delineated by D138^3.32, Y139^3.33 and M142^3.36.

Figure 3.

Putative binding mode for compounds a) 1, b) 4d, c) 5a and d) 5m.

For each of the compounds 4a–h and 5a–n, a docked pose analogous to 1 was obtained for the KOR, although the various substituents at the C2 position bound in somewhat different orientations. These varying orientations were able to provide insights regarding the possible interaction modes of the compounds at the KOR. For the compounds reported here, the electrophilic Michael acceptor atom of the C2 substituent was never closer than ~7 Å from the S^γ atom of C315 in any of the solutions (taking into account C315 side chain rotameric flexibility). As discussed in Section 2.1.2 above, there may be multiple reasons for this, providing a rationale for why these compounds do not form covalently-bound adducts with KOR.

The putative binding mode for the most affine analog from the conjugated/nonconjugated series (4d) is shown in Figure 3b and is analogous to the putative binding mode for salvinorin A (1), except that the olefinic side chain of the C2 substituent is able to effectively fill the small hydrophobic subpocket formed by D138^3.32, Y139^3.33 and M142^3.36 (Figure S2). The model shows how the presence of an intervening methyl group between the ester and olefin functionalities, as in 4d, allows the olefinic portion to be directed into the small subpocket. Other analogs in the conjugated and non-conjugated series were not predicted to bind completely into the subpocket, presumably due to steric incompatibilities with methyl groups in the R₂ and R₃ positions (See Scheme 1) of the C2 substituent. Interestingly, methyl substitution at the R₁ position is tolerated (see 4b and 4h) and the model shows that there is a region of steric tolerance at this locus.

The added steric bulk associated with compounds of the aromatic series 5a–n placed the C2 substituent beyond the Y139^3.33/M142^3.36 subpocket into a hydrophobic pocket bounded by side chains from TM helices 3, 5, and 6 including M142^3.36, K227^5.39, V230^5.42 and I294^6.55. The putative binding mode for the cinnamoyl analog 5a, the most potent and efficacious of the tested compounds in the aromatic series at KOR, is shown in Figure 3c. For 5a, the olefin portion serves as a hydrophobic linker, allowing the phenyl group to effectively engage the hydrophobic regions of its putative binding site.

The addition of small substituents to the phenyl ring (either electron donating or withdrawing) resulted in significantly decreased affinity for KOR, and the model suggests that this is due to steric intolerance within the binding site. Phenyl substitution at the ortho position is the most well tolerated, and the models show that there is a region of steric tolerance at that position where a substituent may be placed. For analogs with meta and para substitutions, no docked solutions were found that placed the aromatic portion directly into the aforementioned site. The trisubstituted analog 5f has no significant affinity for any of the three opioid receptors, probably due one again to steric intolerance within the orthosteric binding site.

The remarkable observation that the 3-pyridyl analog 5m has no affinity for the KOR suggests that the polarity introduced by the heterocyclic nitrogen atom cannot be tolerated in the KOR binding site occupied by 5a. The putative binding site for the phenyl group in Figure 3c is indeed very hydrophobic. On the other hand, 5m has substantial MOR affinity (K_i = 334 nM). A comparison of the residues in the putative aromatic binding pocket of KOR and MOR reveals that one of the amino acids is different (KOR, I294^6.55; MOR, V302^6.55). However, this conservative change does not explain the KOR/MOR selectivity of 5m. Thus, an alternate binding mode was proposed for 5m in MOR; this model is shown in Figure 3d. Here, the pyridine nitrogen of 5m hydrogen bonds with the non-conserved N127^2.63 (position 2.63: KOR = V; MOR = N; DOR = K) and the furan substituent is sandwiched via π-stacking interactions between Y326^7.43 and W293^6.48 (mMOR numbering). The isosteric 3-furanyl analog 5n has excellent MOR affinity (K_i = 18 nM) and may interact in the same way, although in this case the interaction may be of a more hydrophobic/aromatic nature, as thiophene is a poorer hydrogen bond acceptor that pyridine. The greater hydrophobicity of thiophene compared to pyridine may also explain the greater affinity of 5n for the KOR (K_i = 304 nM), and as such may also adopt a binding mode analogous to that proposed for 5a in Figure 3c.

3. Conclusion

A series of potential Michael acceptor-type analogues of salvinorin A were synthesized in an effort to explore the effects of C-2 substitution at opioid receptors. Binding assay results showed that none of the synthesized analogues were able to irreversibly bind to the κ-opioid receptor. Aliphatic conjugated and non-conjugated series of analogues showed high binding affinity to the κ-opioid receptor. Introduction of a cinnamoyl and thiophenylacryloyl group at C-2 had significant impact on affinity, resulting in discovery of new ligands with dual affinity to κ- and μ-opioid receptors. Molecular modeling techniques were employed to determine potential binding modes for the compounds. According to molecular modeling data, the electrophilic center of the studied compounds was distant from the S^γ atom of C315, with a distance of 7 Å or more, confirming the lack of close proximity necessary for covalent binding.

4. Materials and methods

4.1. Experimental section

All commercially available reagents were used without further purification unless otherwise noted. The reactions were performed under an argon atmosphere in anhydrous dichloromethane (DCM) purchased from Sigma-Aldrich. The ¹H NMR spectra were recorded on a Bruker Avance–400 spectrometer using CDCl₃ as solvent, δ values are in ppm and coupling constants in Hz. Melting points were determined on an MEL-TEMP^® 3.0 apparatus and are uncorrected. Thin-layer chromatography (TLC) was performed on 250 μm layer (Whatman PE SIL G/UV silica gel) polyester plates using n-hexane/EtOAc (1:1) as solvent. TLC spots were visualized with anisaldehyde/H₂SO₄ in methanol. Column chromatography was performed on silica gel (230×400 mesh) purchased from Sorbent Technologies (Atlanta, GA). Analytical HPLC was performed on a Waters 2487 apparatus with dual λ absorbance detector system using a Phenomenex Luna-C₁₈ column (4.6×250 mm, 5 μm) with gradient elution at a flow rate 0.5–1.0 mL/min. Isolation of salvinorin A was performed according to the earlier reported procedure [30].

4.2. Synthesis and chemistry

4.2.1. General Procedure A: The procedure for the synthesis of compounds 4a-4h

Compound 3 (15 mg, 1 eq) and triethylamine (1.5–3 eq) were dissolved in DCM (3 mL). An appropriate acryloyl chloride (1.5–3 eq) was added, and the reaction mixture was stirred for 3 h. After TLC indicated completion of the reaction, the mixture was quenched with water and the organic layer separated. The organic phase was washed with dilute aqueous HCl (0.01 mol/L, 2 mL) followed by saturated NaHCO₃ (2 mL). The organic layer was dried over anhydrous Na₂SO₄, evaporated, and the residue was purified by column chromatography (SiO₂; eluent: n-hexane/EtOAc) to obtain the target product.

4.2.2. Genera8/11/2014l Procedure B: The procedure for the synthesis of compounds 5a-5n

To a solution of compound 3 (15 mg, 1 eq) in DCM (3 mL), a catalytic amount of 4-(dimethylamino)pyridine (DMAP) and 1,4-dicyclohexylcarbodimide (DCC) (3 eq), and an appropriate cinnamic or heterocyclic acrylic acid (3 eq) were added. The mixture was stirred at room temperature for 2–5 h. After TLC indicated completion of the reaction, the mixture was concentrated under reduced pressure and the residue was purified by column chromatography (SiO₂; eluent: n-hexane/EtOAc) to yield the target product.

2-O-Acryloylsalvinorin B (4a)

Compound 4a was synthesized from 3 and acryloyl chloride (procedure A) to afford 13 mg (73%) of 4a as a white solid, mp 169–171 °C; ¹H NMR (400 MHz, CDCl₃): δ 7.41 (s, 1H), 7.39 (s, 1H), 6.49 (d, J = 16.4 Hz, 1H), 6.38 (s, 1H), 6.20 (dd, J = 8.0, 16.4 Hz, 1H), 5.93 (d, J = 8.0 Hz, 1H), 5.52 (dd, J = 5.2, 11.4 Hz, 1H), 5.23 (dd, J = 7.6, 12.6 Hz, 1H), 3.73 (s, 3H), 2.78 (dd, J = 8.6, 8.6 Hz, 1H), 2.54 (dd, J = 5.2, 13.4 Hz, 1H), 2.37–2.33 (m, 2H), 2.22 (s, 1H), 2.19–2.10 (m, 2H), 1.80 (ddd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.68–1.56 (m, 3H), 1.45 (s, 3H), 1.15 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 201.84, 171.75, 171.21, 164.71, 143.57, 139.30, 132.29, 127.36, 125.21, 108.40, 75.21, 72.01, 63.91, 53.67, 51.91, 51.37, 43.33, 42.06, 38.08, 35.06, 30.75, 17.93, 16.17, 15.18. HRESIMS (m/z): [M+H]⁺ calculated for C₂₄H₂₉O₈, 445.1784; found, 445.1761. HPLC t_R = 10.974 min; purity = 98.69%.

2-O-Methacryloylsalvinorin B (4b)

Compound 4b was synthesized from 3 and methacryloyl chloride (procedure A) to obtain 16 mg (70%) of 4b as a white solid, mp 196–198 °C. The ¹H NMR spectrum of 4b was consistent with that previously reported.[21] ¹³C NMR (100 MHz, CDCl₃): δ 201.97, 171.67, 171.24, 166.15, 143.61, 139.47, 135.30, 126.88, 125.12, 108.45, 75.23, 72.03, 63.95, 53.50, 51.96, 51.34, 43.19, 42.08, 38.10, 35.43, 30.84, 18.19, 16.36, 15.16. HRESIMS (m/z): [M+H]⁺ calculated for C₂₅H₃₁O₈, 459.1941; found, 459.1935. HPLC t_R = 19.306 min; purity = 99.01%.

2-O-(3′,3′-Dimethylacryloyl)salvinorin B (4c)

Compound 4c was synthesized from 3 and 3,3–dimethylacryloyl chloride (procedure A) to afford 9 mg (51%) of 4c as a white solid, mp 183–185 °C; ¹H NMR (400 MHz, CDCl₃): δ 7.41 (s, 1H), 7.38 (s, 1H), 6.38 (s, 1H), 5.79 (s, 1H), 5.51 (dd, J = 5.2, 11.5 Hz, 1H), 5.20 (dd, J = 7.6, 12.6 Hz, 1H), 3.71 (s, 3H), 2.78 (dd, J = 8.6, 8.6 Hz, 1H), 2.51 (dd, J = 5.2, 13.4 Hz, 1H), 2.32–2.28 (m, 2H), 2.20 (s, 1H), 2.16 (s, 3H), 2.13–2.07 (m, 2H), 1.92 (s, 3H), 1.80 (ddd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.67–1.55 (m, 3H), 1.44 (s, 3H), 1.14 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 202.33, 171.68, 171.23, 164.99, 159.16, 143.64, 139.49, 125.13, 114.85, 108.38, 73.96, 71.60, 63.87, 53.34, 51.58, 51.30, 43.07, 41.76, 37.90, 35.33, 30.93, 27.35, 20.46, 18.13, 16.29, 14.99. HRESIMS (m/z): [M+H]⁺ calculated for C₂₆H₃₃O₈, 473.2097; found, 473.2063. HPLC t_R = 21.211 min; purity = 98.91%.

2-O-(3′-Methylbut-3′-enoyl)salvinorin B (4d)

Compound 4d (8.5 mg, 49%) was obtained as a by-product from the reaction of 3 and 3,3-dimethylacryloyl chloride (procedure A) as a white solid, mp 168–170 °C; ¹H NMR (400 MHz, CDCl₃): δ 7.41 (s, 1H), 7.39 (s, 1H), 6.38 (s, 1H), 5.51 (dd, J = 5.2, 11.6 Hz, 1H), 5.16 (dd, J = 7.6, 12.6 Hz, 1H), 4.94 (s, 1H), 4.90 (s, 1H), 3.71 (s, 3H), 3.15 (s, 2H), 2.76 (dd, J = 8.6, 8.6 Hz, 1H), 2.52 (dd, J = 5.2, 13.4 Hz, 1H), 2.30–2.28 (m, 2H), 2.21 (s, 1H), 2.16–2.07 (m, 2H), 1.85 (s, 3H), 1.82 (ddd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.68–1.56 (m, 3H), 1.44 (s, 3H), 1.14 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 201.70, 171.52, 171.07, 170.38, 143.75, 139.53, 137.99, 125.21, 115.17, 108.49, 75.13, 71.80, 63.79, 53.50, 51.94, 51.32, 43.20, 42.90, 42.07, 38.17, 35.43, 30.73, 22.46, 18.16, 16.40, 15.10. HRESIMS (m/z): [M+Na]⁺ calculated for C₂₆H₃₂O₈Na, 495.1996; found, 495.1959. HPLC t_R = 11.457 min; purity = 97.54 %.

2-O-Crotonoylsalvinorin B (4e)

Compound 4e was synthesized from 3 and crotonyl chloride (procedure A) to obtain 8 mg (49%) of 4e as a white solid, mp 163–165 °C; ¹H NMR (400 MHz, CDCl₃): δ 7.41 (s, 1H), 7.39 (s, 1H), 7.13–7.04 (m, 1H), 6.38 (s, 1H), 5.93 (d, J = 7.8 Hz, 1H), 5.51 (dd, J = 5.2, 11.5 Hz, 1H), 5.22 (dd, J = 7.6, 12.6 Hz, 1H), 3.73 (s, 3H), 2.79 (dd, J = 8.6, 8.6 Hz, 1H), 2.51 (dd, J = 5.2, 13.4 Hz, 1H), 2.35–2.28 (m, 2H), 2.25 (s, 1H), 2.13–2.07 (m, 2H), 1.91 (d, J = 7.0 Hz, 3H), 1.80 (ddd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.65–1.55 (m, 3H), 1.45 (s, 3H), 1.12 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 202.13, 171.62, 171.23, 165.00, 146.67, 143.64, 139.48, 125.22, 121.54, 108.45, 74.79, 72.00, 63.84, 53.43, 51.91, 51.27, 43.19, 42.06, 38.09, 35.41, 30.84, 18.11, 16.39, 15.14. HRESIMS (m/z): [M+H]⁺ calculated for C₂₅H₃₁O₈, 459.1941; found, 459.1928. HPLC t_R = 15.681 min; purity = 98.81%.

2-O-(3′-Butenoyl)salvinorin B (4f)

Compound 4f (9 mg, 51%) was obtained as by-product from the reaction of 3 and crotonyl chloride (procedure A) as a white solid, mp 157–159 °C; ¹H NMR (400 MHz, CDCl₃): δ 7.43 (s, 1H), 7.40 (s, 1H), 6.50–6.44 (m, 1H), 6.39 (s, 1H), 5.92 (s, 1H), 5.90 (s, 1H), 5.54 (dd, J = 5.2, 11.5 Hz, 1H), 5.22 (dd, J = 7.6, 12.6 Hz, 1H), 3.74 (s, 3H), 3.40–3.31(m, 2H), 2.78 (dd, J = 8.6, 8.6 Hz, 1H), 2.53 (dd, J = 5.2, 13.4 Hz, 1H), 2.36–2.29 (m, 2H), 2.21 (s, 1H), 2.17–1.97 (m, 2H), 1.81 (ddd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.65–1.56 (m, 3H), 1.47 (s, 3H), 1.14 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 202.43, 171.65, 171.54, 170.36, 143.69, 139.48, 137.96, 125.23, 115.23, 108.42, 75.13, 72.00, 64.03, 53.66, 51.95, 51.36, 43.41, 42.88, 42.12, 38.33, 35.45, 30.94, 18.14, 16.40, 15.16. HRESIMS (m/z): [M+Na]⁺ calculated for C₂₅H₃₀O₈Na, 481.1839; found, 481.1804. HPLC t_R = 23.537 min; purity = 98.88%.

2-O-(2′,3′-Dimethylacryloyl)salvinorin B (4g)

Compound 4g was synthesized from 3 and 2,3–dimethylacryloyl chloride (procedure A) to afford 8.8 mg (49%) of 4g as a white solid, mp 161–163 °C; ¹H NMR (400 MHz, CDCl₃): δ 7.41 (s, 1H), 7.40 (s, 1H), 7.14– 7.05 (m, 2H), 6.38 (s, 1H), 5.52 (dd, J = 5.2, 11.5 Hz, 1H), 5.22 (dd, J = 7.6, 12.6 Hz, 1H), 3.74 (s, 3H), 2.77 (dd, J = 8.6, 8.6 Hz, 1H), 2.55 (dd, J = 5.2, 13.4 Hz, 1H), 2.34– 2.28 (m, 2H), 2.20 (s, 1H), 2.16–2.07 (m, 2H), 1.92 (d, J = 5.6 Hz, 3H), 1.81 (ddd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.79 (s, 3H), 1.67–1.56 (m, 3H), 1.46 (s, 3H), 1.14 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 202.07, 171.78, 170.84, 165.22, 146.69, 143.52, 139.72, 125.02, 121.56, 108.27, 75.28, 72.25, 63.78, 53.32, 52.26, 51.04, 43.55, 41.82, 35.17, 30.71, 29.52, 27.82, 23.57, 17.86, 16.22, 14.54. HRESIMS (m/z): [M+H]⁺ calculated for C₂₆H₃₃O₈, 473.2097; found, 473.2059. HPLC t_R = 17.950 min; purity = 98.08%.

2-O-(2′-Methyl-3′-butenoyl)salvinorin B (4h)

Compound 4h (9.5 mg, 51%) was obtained as a by-product from the reaction of 3 and 2,3-dimethylacryloyl chloride (procedure A) as a white solid, mp 168–170 °C; ¹H NMR (400 MHz, CDCl₃): δ 7.40 (s, 1H), 7.38 (s, 1H), 6.37 (s, 1H), 6.00–5.88 (m, 1H), 5.48 (dd, J = 5.2, 11.6 Hz, 1H), 5.23–5.11 (m, 3H), 3.71 (s, 3H), 3.35–3.25 (m, 1H), 2.76 (dd, J = 8.6, 8.6 Hz, 1H), 2.49 (dd, J = 5.2, 13.4 Hz, 1H), 2.35–2.26 (m, 2H), 2.21 (s, 1H), 2.16–2.05 (m, 2H), 1.81 (dd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.67–1.54 (m, 3H), 1.43 (s, 3H), 1.34 (d, J = 5.0 Hz, 3H), 1.11 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 201.77, 173.47, 171.56, 171.17, 143.59, 139.53, 136.56, 125.08, 116.37, 108.49, 74.78, 71.96, 63.88, 53.70, 51.93, 51.25, 43.24, 41.98, 38.12, 35.36, 30.69, 18.17, 16.82, 16.41, 15.12. HRESIMS (m/z): [M+H]⁺ calculated for C₂₆H₃₃O₈, 473.5275; found, 473.5254. HPLC t_R = 30.242 min; purity = 98.97%.

2-O-Cinnamoylsalvinorin B (5a)

According to general procedure B, the title compound (24 mg, 92%) was obtained as a white powder from 3 and cinnamic acid. mp 202–204 °C; ¹H NMR (400 MHz, CDCl₃): δ 7.77 (d, J = 16.0 Hz, 1H), 7.55–7.53 (m, 2H), 7.42– 7.39 (m, 5H), 6.52 (d, J = 16.0 Hz, 1H), 6.38 (s, 1H), 5.53 (dd, J = 5.2, 11.4 Hz, 1H), 5.30 (dd, J = 7.6, 12.6 Hz, 1H), 3.74 (s, 3H), 2.80 (dd, J = 8.6, 8.6 Hz, 1H), 2.56 (dd, J = 5.2, 13.4 Hz, 1H), 2.42–2.37 (m, 2H), 2.23 (s, 1H), 2.20–2.08 (m, 2H), 1.82 (ddd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.72–1.59 (m, 3H), 1.47 (s, 3H), 1.15 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 202.23, 171.53, 171.13, 165.61, 146.46, 143.64, 139.46, 134.11, 130.64, 129.01, 128.18, 125.24, 116.66, 108.45, 75.14, 71.99, 64.00, 53.56, 51.97, 51.41, 43.21, 42.21, 38.18, 35.48, 30.93, 18.16, 16.45, 15.20. HRESIMS (m/z): [M+H]⁺ calculated for C₃₀H₃₃O₈, 521.2097; found, 521.2069. HPLC t_R = 29.642 min; purity = 99.14%.

2-O-(2″-Methoxycinnamoyl)salvinorin B (5b)

According to general procedure B, the compound 5b (13 mg, 63%) was obtained as a white solid from 3 and 2-methoxycinnamic acid. mp 153–155 °C; ¹H NMR (400 MHz, CDCl₃): δ 8.06 (d, J = 16.0 Hz, 1H), 7.52 (d, J = 8.0 Hz, 1H), 7.42–7.34 (m, 3H), 6.96–6.91 (m, 2H), 6.63 (d, J = 16.0 Hz, 1H), 6.39 (s, 1H), 5.51 (dd, J = 5.2, 11.4 Hz, 1H), 5.30 (dd, J = 7.6, 12.6 Hz, 1H), 3.88 (s, 3H), 3.73 (s, 3H), 2.80 (dd, J = 8.6, 8.6 Hz, 1H), 2.54 (dd, J = 5.2, 13.4 Hz, 1H), 2.41–2.36 (m, 2H), 2.27 (s, 1H), 2.17–2.09 (m, 2H), 1.82 (ddd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.68–1.56 (m, 3H), 1.45 (s, 3H), 1.14 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 202.22, 171.70, 171.27, 166.35, 158.45, 143.69, 141.85, 139.48, 131.90, 129.21, 125.26, 123.01, 120.73, 117.22, 111.20, 108.50, 75.06, 72.08, 63.90, 60.39, 55.41, 53.61, 51.99, 51.34, 43.26, 42.11, 38.19, 35.45, 30.92, 18.14, 16.42, 15.24. HRESIMS (m/z): [M+H]⁺ calculated for C₃₁H₃₅O₉, 551.2203; found, 551.2235. HPLC t_R = 32.453 min; purity = 97.98%.

2-O-(4″-Methoxycinnamoyl)salvinorin B (5c)

According to general procedure B, the compound 5c (15 mg, 65%) was yielded as a white solid from 3 and 4-methoxycinnamic acid. mp 191–193 °C; ¹H NMR (400 MHz, CDCl₃): δ 7.72 (d, J = 16.0 Hz, 1H), 7.49 (d, J = 8.0 Hz, 1H), 7.42 (s, 1H), 7.40 (s, 1H), 6.91 (d, J = 8.0 Hz, 2H), 6.41 (d, J = 16.0 Hz, 1H), 6.37 (s, 1H), 5.52 (dd, J = 5.2, 11.4 Hz, 1H), 5.30 (dd, J = 7.6, 12.6 Hz, 1H), 3.85 (s, 3H), 3.74 (s, 3H), 2.81 (dd, J = 8.6, 8.6 Hz, 1H), 2.54 (d, J = 5.2, 13.4 Hz, 1H), 2.41–2.37 (m, 2H), 2.24 (s, 1H), 2.19–2.08 (m, 2H), 1.83 (ddd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.72–1.58 (m, 3H), 1.47 (s, 3H), 1.16 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 202.25, 171.68, 171.24, 166.10, 161.63, 146.17, 143.69, 139.41, 129.89, 126.93, 125.13, 114.38, 114.06, 108.47, 74.92, 72.08, 64.10, 55.38, 53.65, 51.96, 51.42, 43.40, 42.10, 38.19, 35.49, 30.94, 18.19, 16.44, 15.23. HRESIMS (m/z): [M+Na]⁺ calculated for C₃₁H₃₅O₉Na, 573.2203; found, 573.2169. HPLC t_R = 30.539 min; purity = 96.13%.

2-O-(2″,5″-Dimethoxycinnamoyl)salvinorin B (5d)

According to general procedure B, the title compound 5d (13 mg, 60%) was afforded as a white solid from 3 and 2,5-dimethoxycinnamic acid. mp 217–219 °C; ¹H NMR (400 MHz, CDCl₃): δ 8.05 (d, J = 16.4 Hz, 1H), 7.43 (s, 1H), 7.40 (s, 1H), 7.06 (s, 1H), 6.95 (d, J = 8.0 Hz, 1H), 6.92 (d, J = 8.0 Hz, 1H), 6.60 (d, J = 16.4 Hz, 1H), 6.40 (s, 1H), 5.54 (dd, J = 5.2, 11.4 Hz, 1H), 5.32 (dd, J = 7.6, 12.6 Hz, 1H), 3.85 (s, 3H), 3.80 (s, 3H), 3.75 (s, 3H), 2.82 (dd, J = 8.6, 8.6 Hz, 1H), 2.54 (dd, J = 5.2, 13.4 Hz, 1H), 2.43–2.37 (m, 2H), 2.24 (s, 1H), 2.19–2.08 (m, 2H), 1.80 (ddd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.71–1.58 (m, 3H), 1.48 (s, 3H), 1.16 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 202.17, 171.62, 171.18, 166.18, 153.45, 153.02, 143.68, 141.64, 139.43, 125.22, 123.58, 117.67, 117.26, 113.32, 112.44, 108.40, 75.00, 72.08, 64.11, 56.03, 55.78, 53.67, 51.96, 51.41, 42.26, 38.21, 35.53, 33.94, 30.97, 18.20, 16.45, 15.21. HRESIMS (m/z): [M+H]⁺ calculated for C₃₂H₃₇O₁₀, 581.2308; found, 581.2287. HPLC t_R = 29.079 min; purity = 98.12%.

2-O-(2″,4″-Dimethoxycinnamoyl)salvinorin B (5e)

According to general procedure B, the compound 5e (14 mg, 61%) was obtained as a white solid from 3 and 2,4-dimethoxycinnamic acid. mp 178–180 °C; ¹H NMR (400 MHz, CDCl₃): δ 7.98 (d, J = 16.0 Hz, 1H), 7.46–7.40 (m, 3H), 6.55–6.44 (m, 3H), 6.39 (s, 1H), 5.52 (dd, J = 5.2, 11.4 Hz, 1H), 5.31 (dd, J = 7.6, 12.6 Hz, 1H), 3.87 (s, 3H), 3.84 (s, 3H), 3.74 (s, 3H), 2.81 (dd, J = 8.6, 8.6 Hz, 1H), 2.54 (dd, J = 5.2, 13.4 Hz, 1H), 2.41–2.36 (m, 2H), 2.24 (s, 1H), 2.19–2.08 (m, 2H), 1.83 (ddd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.71–1.58 (m, 3H), 1.46 (s, 3H), 1.15 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 202.51, 171.80, 171.32, 166.67, 163.04, 159.96, 143.73, 141.65, 139.36, 130.96, 125.13, 116.41, 114.36, 108.38, 105.32, 99.39, 74.83, 72.08, 64.07, 55.56, 53.76, 51.91, 51.39, 43.31, 42.02, 38.19, 35.51, 30.94, 18.19, 16.43, 15.16. HRESIMS (m/z): [M+H]⁺ calculated for C₃₂H₃₆O₁₀Na, 603.2342; found, 603.2389. HPLC t_R = 23.628 min; purity = 98.69%.

2-O-(3″,4″,5″-Trimethoxycinnamoyl)salvinorin B (5f)

According to the general procedure B, the compound 5f (16 mg, 64%) was yielded as a white solid from 3 and 3, 4,5-trimethoxycinnamic acid. mp 222–224 °C; ¹H NMR (400 MHz, CDCl₃): δ 7.68 (d, J = 16.0 Hz, 1H), 7.42 (s, 1H), 7.39 (s, 1H), 6.77 (s, 2H), 6.43 (d, J = 16.0 Hz, 1H), 6.39 (s, 1H), 5.53 (dd, J = 5.2, 11.4 Hz, 1H), 5.30 (dd, J = 7.6, 12.6 Hz, 1H), 3.88 (s, 9H), 3.73 (s, 3H), 2.81 (dd, J = 8.6, 8.6 Hz, 1H), 2.52 (dd, J = 5.2, 13.4 Hz, 1H), 2.41–2.37 (m, 2H), 2.24 (s, 1H), 2.18–2.09 (m, 2H), 1.82 (ddd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.73–1.58 (m, 3H), 1.46 (s, 3H), 1.15 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 202.22, 171.61, 171.07, 165.66, 153.38, 146.32, 143.71, 140.31, 139.29, 129.61, 125.26, 115.93, 108.40, 105.45, 75.07, 72.04, 60.88, 56.16, 53.55, 51.99, 51.34, 43.31, 42.11, 38.19, 35.51, 30.89, 18.18, 16.41, 15.22. HRESIMS (m/z): [M+H]⁺ calculated for C₃₃H₃₉O₁₁, 611.2414; found, 611.2403. HPLC t_R = 28.904 min; purity = 99.08%.

2-O-(3″,4″-Dioxymethylenecinnamoyl)salvinorin B (5g)

According to the general procedure B, the title compound (13 mg, 61%) was yielded as a white solid from 3 and 3,4-dioxymethylenecinnamic acid. mp 177–179 °C; ¹H NMR (400 MHz, CDCl₃): δ 7.67 (d, J = 16.0 Hz, 1H), 7.42 (s, 1H), 7.39 (s, 1H), 7.08–7.01 (m, 2H), 6.82 (d, J = 8.0 Hz, 2H), 6.39 (s, 1H), 6.34 (d, J = 16.0 Hz, 1H), 6.02 (s, 2H), 5.53 (dd, J = 5.2, 11.4 Hz, 1H), 5.29 (dd, J = 7.6, 12.6 Hz, 1H), 3.74 (s, 3H), 2.82 (dd, J = 8.6, 8.6 Hz, 1H), 2.53 (dd, J = 5.2, 13.4 Hz, 1H), 2.41–2.37 (m, 2H), 2.24 (s, 1H), 2.20–2.08 (m, 2H), 1.83 (ddd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.72–1.58 (m, 3H), 1.46 (s, 3H), 1.15 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 202.27, 171.60, 171.17, 165.88, 149.89, 148.40, 146.11, 143.62, 139.45, 128.56, 125.28, 124.90, 114.43, 108.58, 106.57, 101.66, 74.86, 72.64, 64.08, 53.76, 52.05, 51.40, 43.38, 42.15, 38.19, 35.48, 30.81, 18.16, 16.44, 15.27. HRESIMS (m/z): [M+H]⁺ calculated for C₃₁H₃₃O₁₀, 565.2029; found, 565.2057. HPLC t_R = 23.401 min; purity = 98.06%.

2-O-(2″-Nitrocinnamoyl)salvinorin B (5h)

According to general procedure B, the title compound 5h (16 mg, 71%) was afforded as a white solid from 3 and 2-nitrocinnamic acid. mp 195–197 °C; ¹H NMR (400 MHz, CDCl₃): δ 8.21 (d, J = 16.0 Hz, 1H), 8.06 (d, J = 8.0 Hz, 1H), 7.68–7.66 (m, 2H), 7.59–7.56 (m, 1H), 7.42 (s, 1H), 7.39 (s, 1H), 6.46 (d, J = 16.0 Hz, 1H), 6.39 (s, 1H), 5.53 (dd, J = 5.2, 11.4 Hz, 1H), 5.31 (dd, J = 7.6, 12.6 Hz, 1H), 3.74 (s, 3H), 2.82 (dd, J = 8.6, 8.6 Hz, 1H), 2.52 (dd, J = 5.2, 13.4 Hz, 1H), 2.42–2.36 (m, 2H), 2.26 (s,1H), 2.16–2.09 (m, 2H), 1.82 (ddd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.70–1.56 (m, 3H), 1.46 (s, 3H), 1.15 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 201.82, 171.47, 171.07, 164.52, 148.36, 143.73, 141.55, 139.48, 133.56, 130.61, 130.19, 129.17, 125.17, 121.83, 108.44, 75.45, 72.07, 64.07, 53.57, 52.00, 51.38, 43.31, 42.19, 38.18, 35.45, 30.87, 18.18, 16.47, 15.23. HRESIMS (m/z): [M+H]⁺ calculated for C₃₀H₃₂NO₁₀, 566.1951; found, 566.1989. HPLC t_R = 21.901 min; purity = 96.93%.

2-O-(4″-Nitrocinnamoyl)salvinorin B (5i)

According to the general procedure B, the title compound 5i (13 mg, 59%) was yielded as a white solid from 3 and 4-nitrocinnamic acid. mp 217–219 °C; ¹H NMR (400 MHz, CDCl₃): δ 8.27 (d, J = 8.0 Hz, 1H), 7.80 (d, J = 16.0 Hz, 1H), 7.70 (d, J = 8.0 Hz, 2H), 7.43 (s, 1H), 7.41 (s, 1H), 6.65 (d, J = 16.0 Hz, 1H), 6.40 (s, 1H), 5.54 (dd, J = 5.2, 11.4 Hz, 1H), 5.31 (dd, J = 7.6, 12.6 Hz, 1H), 3.75 (s, 3H), 2.82 (dd, J = 8.6, 8.6 Hz, 1H), 2.53 (dd, J = 5.2, 13.4 Hz, 1H), 2.44–2.38 (m, 2H), 2.25 (s, 1H), 2.18–2.09 (m, 2H), 1.83 (ddd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.72–1.56 (m, 3H), 1.47 (s, 3H), 1.16 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 201.81, 171.51, 171.06, 164.83, 148.65, 143.75, 143.19, 140.66, 139.37, 128.66, 125.14, 124.16, 120.78, 108.41, 75.52, 71.99, 64.16, 53.60, 52.05, 51.41, 43.36, 42.15, 38.18, 35.42, 30.78, 18.16, 16.48, 15.16. HRESIMS (m/z): [M+Na]⁺ calculated for C₃₀H₃₁NO₁₀Na, 588.1987; found, 588.1951. HPLC t_R = 33.586 min; purity = 98.73%.

2-O-(2″-Trifluoromethylcinnamoyl)salvinorin B (5j)

According to general procedure B, the title compound 5j (14 mg, 62%) was obtained as a white solid from 3 and 2-trifluoromethylcinnamic acid. mp 208–210 °C; ¹H NMR (400 MHz, CDCl₃): δ 8.14 (d, J = 16.0 Hz, 1H), 7.75–7.70 (m, 2H), 7.58–7.55 (m, 1H), 7.54–7.52 (m, 1H), 7.42 (s, 1H), 7.40 (s, 1H), 6.51 (d, J = 16.0 Hz, 1H), 6.39 (s, 1H), 5.52 (dd, J = 5.2, 11.4 Hz, 1H), 5.30 (dd, J = 7.6, 12.6 Hz, 1H), 3.74 (s, 3H), 2.81 (dd, J = 8.6, 8.6 Hz, 1H), 2.52 (dd, J = 5.2, 13.4 Hz, 1H), 2.42–2.37 (m, 2H), 2.27 (s, 1H), 2.19 –2.09 (m, 2H), 1.82 (ddd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.70–1.58 (m, 3H), 1.46 (s, 3H), 1.14 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 201.90, 171.58, 171.15, 164.82, 143.62, 141.65, 139.55, 132.97, 132.10, 129.83, 127.97, 126.24, 125.20, 121.01, 108.50, 75.34, 72.08, 64.02, 53.54, 51.89, 51.33, 43.30, 42.15, 38.11, 35.53, 30.80, 18.16, 16.43, 15.14. HRESIMS (m/z): [M+H]⁺ calculated for C₃₁H₃₂F₃O₈, 589.1971; found, 589.1996. HPLC t_R = 42.336 min; purity = 98.04%.

2-O-(4″-Trifluoromethylcinnamoyl)salvinorin B (5k)

According to general procedure B, the title compound 5k (12 mg, 60%) was obtained as a white solid from 3 and 4-trifluoromethylcinnamic acid. mp 148–150 °C; ¹H NMR (400 MHz, CDCl₃): δ 7.77 (d, J = 16.0 Hz, 1H), 7.68–7.63 (m, 4H), 7.42 (s, 1H), 7.40 (s, 1H), 6.60 (d, J = 16.0 Hz, 1H), 6.39 (s, 1H), 5.54 (dd, J = 5.2, 11.4 Hz, 1H), 5.31 (dd, J = 7.6, 12.6 Hz, 1H), 3.75 (s, 3H), 2.82 (dd, J = 8.6, 8.6 Hz, 1H), 2.53 (dd, J = 5.2, 13.4 Hz, 1H), 2.42–2.36 (m, 2H,), 2.25 (s, 1H), 2.15–2.05 (m, 2H), 1.82 (ddd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.69–1.58 (m, 3H), 1.47 (s, 3H), 1.16 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 201.83, 171.54, 171.09, 165.22, 144.43, 143.70, 139.45, 137.45, 128.35, 125.23, 119.36, 108.40, 75.32, 71.99, 64.15, 53.52, 52.02, 51.37, 43.42, 42.16, 38.21, 35.52, 30.89, 18.22, 16.46, 15.22. HRESIMS (m/z): [M+Na]⁺ calculated for C₃₁H₃₁F₃O₈Na, 611.1893; found, 611.1867. HPLC t_R = 52.609 min; purity = 97.79%.

2-O-[2″,5″-bis(Trifluoromethyl)cinnamoyl]salvinorin B (5l)

According to general procedure B, the title compound 5l (17 mg, 66%) was afforded as a white solid from 3 and 2,5-di(trifluoromethyl)cinnamic acid. mp 185–187 °C; ¹H NMR (400 MHz, CDCl₃): δ 8.11 (d, J = 16.0 Hz, 1H), 7.98 (s, 1H), 7.86 (d, J = 8.0 Hz, 1H), 7.76 (d, J = 8.0 Hz, 1H), 7.42 (s, 1H), 7.40 (s, 1H), 6.59 (d, J = 16.0 Hz, 1H), 6.39 (s, 1H), 5.52 (dd, J = 5.2, 11.4 Hz, 1H), 5.30 (dd, J = 7.6, 12.6 Hz, 1H), 3.74 (s, 3H), 2.82 (dd, J = 8.6, 8.6 Hz, 1H), 2.52 (dd, J = 5.2, 13.4 Hz, 1H), 2.42–2.37 (m, 2H), 2.27 (s, 1H), 2.21–2.13 (m, 2H), 1.83 (ddd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.71–1.58 (m, 3H), 1.46 (s, 3H), 1.15 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 201.62, 171.46, 171.48, 164.34, 143.71, 140.00, 134.24, 130.83, 128.79, 127.12, 125.20, 123.01, 122.29, 108.43, 75.55, 72.01, 64.10, 53.46, 51.99, 51.34, 43.23, 42.15, 38.15, 35.49, 30.79, 18.15, 16.40, 15.13. HRESIMS (m/z): [M+H]⁺ calculated for C₃₂H₃₁F₆O₈, 657.1818; found, 657.1803. HPLC t_R = 71.301 min; purity = 98.95%.

2-O-[3′-(3″-Pyridyl)acryloyl]salvinorin B (5m)

According to general procedure B, the compound 5m (13 mg, 63%) was obtained as a white solid from 3 and trans 3-(3-pyridyl)acrylic acid. mp 189–191 °C; ¹H NMR (400 MHz, CDCl₃): δ 8.77 (s, 1H), 8.62 (d, J = 8.0 Hz, 1H), 7.88 (d, J = 8.0 Hz, 1H), 7.75 (d, J = 16.0 Hz, ¹H), 7.41–7.37 (m, 2H), 6.60 (d, J = 16.0 Hz, 1H), 6.38 (s, 1H), 5.52 (dd, J = 5.2, 11.4 Hz, 1H), 5.31 (dd, J = 7.6, 12.6 Hz, 1H), 3.74 (s, 3H), 2.82 (dd, J = 8.6, 8.6 Hz, 1H), 2.53 (dd, J = 5.2, 13.4 Hz, 1H), 2.42–2.37 (m, 2H), 2.26 (s, 1H), 2.18–2.11 (m, 2H), 1.82 (ddd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.68–1.57 (m, 3H), 1.46 (s, 3H), 1.15 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 201.93, 171.49, 171.10, 165.11, 150.88, 149.52, 143.61, 142.35, 139.50, 134.67, 129.91, 125.19, 123.86, 119.13, 108.32, 75.37, 72.01, 63.84, 53.41, 51.77, 51.21, 43.35, 41.90, 37.96, 35.49, 30.65, 18.01, 16.29, 15.20. HRESIMS (m/z): [M+H]⁺ calculated for C₂₉H₃₂NO₈, 522.2083; found, 522.2069. HPLC t_R = 10.043 min; purity = 97.11 %.

2-O-[3′-(3″-Thiophenyl)acryloyl]salvinorin B (5n)

According to general procedure B, the compound 5n (15 mg, 67%) was yielded as a white solid from 3 and 3-thiopheneacrylic acid. mp 184–186 °C; ¹H NMR (400 MHz, CDCl₃): δ 7.75 (d, J = 16.4 Hz, 1H), 7.53 (s, 1H), 7.42 (s, 1H), 7.40 (s, 1H), 7.35–7.32 (m, 2H), 6.39 (s, 1H), 6.35 (d, J = 16.4 Hz, 1H), 5.54 (dd, J = 5.2, 11.4 Hz, 1H), 5.29 (dd, J = 7.6, 12.6 Hz, 1H), 3.74 (s, 3H), 2.80 (dd, J = 8.6, 8.6 Hz, 1H), 2.54 (dd, J = 5.2, 13.4 Hz, 1H), 2.42–2.38 (m, 2H), 2.23 (s, 1H), 2.19–2.08 (m, 2H), 1.82 (ddd, J = 3.0, 3.0, 10.4 Hz, 1H), 1.68–1.57 (m, 3H), 1.47 (s, 3H), 1.15 (s, 3H). ¹³C NMR (100 MHz, CDCl₃): δ 202.21, 171.51, 171.14, 165.78, 143.69, 139.78, 139.58, 137.03, 128.73, 127.11, 124.89, 116.55, 108.56, 75.03, 71.82, 64.04, 53.63, 51.97, 51.45, 43.40, 42.15, 35.54, 30.84, 18.25, 16.44, 15.06. HRESIMS (m/z): [M+H]⁺ calculated for C₂₈H₃₁O₈S, 527.1661; found, 527.1637. HPLC t_R = 23.444 min; purity = 97.02 %.

4.3. Pharmacology

4.3.1 Affinity Determinations: HEK293T-KOR Membrane Preparation and Radioligand Binding Assay

HEK293T membrane preparation

Cells stably expressing κ-opioid receptors (HEK293T-KOR) were plated in 15-cm dishes (in DMEM containing 10% FBS) and grown to 90% confluence [10]. Cells were then washed with PBS at pH 7.4, and harvested by scraping into PBS, pH 7.4. Harvested cells were centrifuged at 1,000 × g for 10 min and subsequently hypotonically lysed by resuspension into ice-cold binding buffer (50mM Tris-HCl, 10mM MgCl₂ 0.1mM EDTA, pH 7.4). Membranes were isolated by centrifugation at 21,000 × g for 20 min. The supernatant was removed and the membrane pellets were stored at −80 °C until used for radioligand binding assays.

4.3.2. Radioligand binding assay

Membranes prepared as above were resuspended in binding buffer to 1 μg protein/μL (measured by Bradford assay using BSA as standard), and 50 μL was added to each well of a polypropylene 96-well plate containing 200 μL binding buffer with 0.5 nM [³H]U-69593, and reference or test ligands at various concentrations ranging from 10 pM to 10 μM (final concentrations). After 1.5 h incubation in the dark at room temperature, the reactions were harvested onto 0.3% PEI-soaked Filtermax GF/A filters (Wallac) and washed three times with ice-cold 50 mM Tris, pH 7.4, using a Perkin-Elmer Filtermate 96-well harvester. The filters were subsequently dried and placed on a hot plate (100 °C), and Melitilex-A (Wallac) scintillant was applied. The filters were then removed from the hot plate and allowed to cool. The filters were counted on a Wallac TriLux MicroBeta counter (3 min/well). Residual [³H] U-69593 binding to filtered membranes was plotted as a function of log [ligand] and the data were regressed using the one-site competition model built into Prism 4.0 (GraphPad software).

4.3.3. cAMP Inhibition: KOR-Mediated cAMP Assay

HEK293T cells were co-transfected with plasmids encoding the cAMP biosensor GloSensor-22F (Promega) and hKOR receptor. After 18 h incubation at 37 °C the cells were seeded (20,000 cells/20 μL/well) into white, clear-bottom, 384 well tissue culture plates in HBSS, 10% FBS, 20 mM HEPES, pH 7.4. After a 1- to 2-h recovery, cells were treated with 10 μL of 3× test or reference drug prepared in HBSS, 10% FBS, 20 mM HEPES, pH 7.4. After 30 min, cAMP production was stimulated and detected by treatment with 10 μL of 1,200 nM (4×) isoproterenol in 8% (4×) GloSensor reagent. Luminescence per well per second was read on a Wallac TriLux MicroBeta plate counter. Data were normalized to the isoproterenol response (0%) and the maximal Salvinorin A-induced inhibition thereof (100%) and regressed using the sigmoidal dose-response function built into GraphPad Prism 4.0.

4.4. Molecular Modeling

Molecular modeling investigations were conducted using the SYBYL-X molecular modeling package (version 2.0, 2012, Tripos Associates, St. Louis, MO). GOLDSuite 5.1 (Cambridge Crystallographic Data Centre, Cambridge, UK) was employed to perform automated docking tasks. Default parameters were used unless otherwise noted. The molecular modeling methods employed here are analogous to those previously published (see Wu, et al.,[9] Vardy, et al.[28] and references within).

The structure of the hKOR was prepared for docking by extracting the ‘B’ chain (protein only) from the PDB file (PDB ID = 4DJH) and adding hydrogen atoms, and extracting the JDTic ligand. Five water molecules (1307, 1311, 1313, 1314, 1316) located in the ‘B’ chain binding site were also extracted and saved individually for use in the docking exercises. The structure of the mMOR was prepared for docking in an analogous fashion by extracting the ‘A’ chain from PDB ID = 4DJH. As there is no difference in the amino acid composition of the hMOR and mMOR in the orthosteric binding site, the compounds were docked into the 4DJH structure. The structures of the compounds to be docked were sketched in SYBYL and energy-minimized using the Tripos Force Field (Gasteiger–Hückel charges, distance-dependent dielectric constant = 4.0; non-bonded interaction cutoff = 8 Å; termination criterion = energy gradient < 0.05 kcal/(mol×Å) or 100,000 iterations).

GOLD 5.1 flexible docking was performed without constraints for all ligands. GoldScore was used as the scoring/fitness function. For each receptor, a binding site definition consisting of a 15-Å radius sphere around the D(3.32) C^β atom was employed. Water molecules were toggled on or off as required, with 10 solutions generated for each run. To more fully explore potential binding modes for the docked compounds, the Diverse Solutions option in GOLD was enabled with the cluster size set to 1 and 1.5 Å RMSD. The final receptor–ligand complex for each ligand was chosen interactively by selecting the highest-scoring pose that was consistent with experimentally-derived information about the binding mode of the ligand. The chosen solution(s) were then merged back into the receptor structure, along with any necessary water molecules. Finally, the complexes were energy-minimized in SYBYL using the TFF with the previously-stated parameters. The stereochemical quality of the final models was assessed using PROCHECK.

All computations were performed on quad-core Intel Xeon-based SGI Virtu VS (CentOS 6.1), AMD Opteron-based Hewlett-Packard xw9400 (CentOS 6.1) and 8-core Intel Xeon-based Mac Pro (OS-X 10.6 Snow Leopard) workstations.

Highlights.

• ➢New potential Michael acceptor analogues of salvinorin A were synthesized.
• ➢Compounds were evaluated for binding affinity at κ-, δ-, and μ-opioid receptors.
• ➢Molecular modeling studies describe putative binding modes for the compounds.
• ➢Most compounds have high binding affinity at κ; 5a has dual affinity for κ and μ.
• ➢5a could be developed as a potent CNS or peripheral drug in the near future.

Abbreviations used

DOR

δ-opioid receptor

GPCR

G protein-coupled receptor

KOR

κ-opioid receptor

MOR

μ-opioid receptor

TM

transmembrane

PDB

Protein Data Bank