Molecular Docking Studies of Royleanone Diterpenoids from Plectranthus spp. as P-Glycoprotein Inhibitors

Abstract

The development of multidrug resistance (MDR) is a major cause of failure in cancer chemotherapy. Several abietane diterpenes with antitumoral activities have been isolated from Plectranthus spp. such as 6,7-dehydroroyleanone (DHR, 1) and 7α-acetoxy-6β-hydroxyroyleanone (AHR, 2). Several royleanone derivatives were prepared through hemisynthesis from natural compounds 1 and 2 to achieve a small library of products with enhanced anti-P-glycoprotein activity. Nonetheless, some derivatives tend to be unstable. Therefore, to reason such lack of stability, the electron density based local reactivity descriptors condensed Fukui functions and dual descriptor were calculated for several derivatives of DHR. Additionally, molecular docking and molecular dynamics studies were performed on several other derivatives to clarify the molecular mechanisms by which they may exert their inhibitory effect in P-gp activity. The analysis on local reactivity descriptors was important to understand possible degradation pathways and to guide further synthetic approaches toward new royleanone derivatives. A molecular docking study suggested that the presence of aromatic moieties increases the binding affinity of royleanone derivatives toward P-gp. It further suggests that one royleanone benzoylated derivative may act as a noncompetitive efflux modulator when bound to the M-site. The future generation of novel royleanone derivatives will involve (i) a selective modification of position C-12 with chemical moieties smaller than unsubstituted benzoyl rings and (ii) the modification of the substitution pattern of the benzoyloxy moiety at position C-6.

While cancer is one of the leading causes of death worldwide,¹ the development of multidrug resistance (MDR) is a major cause of failure in traditional cancer chemotherapy.² Development of MDR is often associated with mutations or overexpression of ABC (ATP-binding cassette) family of membrane transport proteins such as ABCB1, also known as MDR1 or P-glycoprotein (P-gp). P-gp is a key player in the MDR by mediating ATP-dependent efflux of many anticancer drugs, decreasing their intracellular concentration and thus conferring resistance to these agents.^3,4 Nevertheless, other events, such as deregulation in the apoptosis signaling, are also accountable for the phenomenon of MDR.⁵

The isolation of new natural products provides novel scaffolds with potential biological activities for cancer therapy. Several Plectranthus species (Lamiaceae) have been used in traditional medicine.⁶ The most commonly reported secondary metabolites in this genus are diterpenoids, mainly abietane quinones. Several abietane diterpenes isolated from Plectranthus spp. have been described as antimicrobial, antifungal, and cytotoxic.⁷⁻¹⁰ Compound 6,7-dehydroroyleanone (DHR, 1, Figure 1) is the main component of P. madagascariensis (Pers.) Benth essential oil and exhibits several interesting bioactivities.¹¹ The cytotoxic diterpene 7α-acetoxy-6β-hydroxyroyleanone (AHR, 2, Figure 1) is the main metabolite of P. grandidentatus extracts.⁸ The reported bioactivities and molecular structures suitable for derivatization make DHR and AHR attractive leads to the drug discovery process.

Figure 1.

6,7-Dehydroroyleanone (DHR, 1), 7α-acetoxy-6β-hydroxyroyleanone (AHR, 2), 6β,7α-dihydroxyroyleanone (3), 7α-acetoxy-6β-(4-chloro)benzoyloxy-12-O-(4-chloro)benzoylroyleanone (4), and 7α-acetoxy-6β-hydroxy-12-O-benzoylroyleanone (5).

In previous works,¹²⁻¹⁵ more than 20 royleanone derivatives were prepared through hemisynthesis from lead molecules 1 and 2. The hemisynthetic derivatives were prepared aiming at building a library of compounds with enhanced cytotoxic potential regarding P-gp inhibition. In the present work, we report our studies on the most promising synthesized derivatives regarding stability and overall reaction yield. The compound 6,7-dihydroroyleanone (3, Figure 1) was obtained by basic hydrolysis of compound 2. Additionally, compounds 7α-acetoxy-6β-(4-chloro)benzoyloxy-12-O-(4-chloro)benzylroyleanone (4, Figure 1) and 7α-acetoxy-6β-hydroxy-12-O-benzoylroyleanone (5, Figure 1) were obtained from benzoylation of the same starting material. The cytotoxic activities of compounds 1, 2, and 3 were assayed in nonsmall cell lung cancer (NCI-H460) and its MDR variant (NCI-H460/R), which overexpress the P-pg. Royleanones 1, 2, and 3 displayed the ability to inhibit both cancer cell lines (NCI-H460 and NCI-H460/R).^8,16 Additionally, in a recent study, the diterpene 4 exhibited promising P-gp inhibition, with highest toxicity in NCI-H460 and NCI-H460/R lines, when compared to the compounds 1, 2, and 3.¹⁵ The anti-Pgp activity was never assessed for royleanone 5.

Several reactions were performed in the natural compounds 1 and 2, and some substrate diversity was tested with the intention of achieving compounds with enhanced cytotoxic potential. Nonetheless, the broad diversity tested, and several derivatives presented stability issues in the isolation process. For instance, all derivatives obtained by microwave-assisted Mitsunobu reaction of 1 and 2 were found to be unstable and, thus, unable to be isolated or further characterized. On the other hand, the hydrogenation, methylation, or benzoylation reactions of 1 or 2 resulted in stable products. Several attempts to tosylate compound 1 led to formation of a single product as judged by TLC, and the presence of 7 was identified by ¹H NMR of crude mixture. However, attempts for its isolation were unsuccessful due to decomposition, although the presence of the desired compound could be confirmed by HRMS.

Therefore, to get insight on the causes of such lack of stability, the electron density based local reactivity descriptors condensed Fukui functions^17,18 and dual descriptor were calculated for the DHR (1), methylated (6), and tosylated (7) derivatives. Herein, Fukui functions have the capability of revealing the presence of nucleophilic and electrophilic regions within a molecule or a given scaffold, thus providing clues on the decomposition pathways that may have been followed by royleanone derivatives. Starting with neutral species, their geometry was first optimized at PBE1PBE/6-31G(d,p) level of theory. The natural bond orbitals of the neutral, cation, and anion species were used for determination of the electrophilic (f_k^-), nucleophilic (f_k⁺), and radical attacks (f_k⁰) functions, and their condensed versions for carbon and oxygen atoms are presented in Table 1. The condensed Fukui functions are calculated according to the following equations:

1

2

3

where k is the numbering of the atomic site, and q_K stands for the total natural population in the neutral (N), anionic (N + 1), and cationic (N – 1) species. The dual descriptor (Δf_k), a more accurate local reactivity descriptor,¹⁹⁻²¹ was obtained in its condensed form according to the following equation:

4

Table 1. Values of Condensed Fukui Functions in DHR (1), Methylated (6), and Tosylated (7) Derivatives Computed from Natural Population Analysis in Gas Phase.

	fk–			fk+			fk0			Δfk
atom	1	6	7	1	6	7	1	6	7	1	6	7
C1	–0.004	–0.004	–0.004	–0.003	–0.003	–0.003	–0.004	–0.003	–0.003	0.000	0.001	0.001
C2	–0.005	–0.005	–0.005	–0.005	–0.005	–0.004	–0.005	–0.005	–0.005	0.000	0.000	0.001
C3	–0.006	–0.006	–0.006	–0.005	–0.005	–0.005	–0.006	–0.006	–0.005	0.000	0.000	0.001
C4	–0.004	–0.004	–0.004	–0.004	–0.004	–0.003	–0.004	–0.004	–0.004	0.000	0.001	0.001
C5	–0.018	–0.018	–0.019	–0.014	–0.013	–0.011	–0.016	–0.015	–0.015	0.004	0.005	0.008
C6	0.140	0.140	0.153	0.124	0.112	0.097	0.132	0.126	0.125	–0.016	–0.028	–0.056
C7	0.061	0.059	0.089	–0.040	–0.039	–0.038	0.011	0.010	0.025	–0.101	–0.098	–0.127
C8	0.018	0.026	0.029	0.098	0.082	0.061	0.058	0.054	0.045	0.080	0.057	0.031
C9	0.134	0.126	0.144	0.041	0.046	0.051	0.087	0.086	0.097	–0.093	–0.080	–0.093
C10	–0.010	–0.009	–0.011	–0.005	–0.005	–0.005	–0.008	–0.007	–0.008	0.005	0.004	0.005
C11	–0.031	–0.029	–0.026	0.116	0.107	0.086	0.043	0.039	0.030	0.147	0.137	0.112
O11′	0.055	0.056	0.057	0.133	0.138	0.133	0.094	0.097	0.095	0.078	0.082	0.076
C12	0.041	0.041	0.032	0.015	0.031	0.050	0.028	0.036	0.041	–0.026	–0.010	0.017
O12′	0.077	0.073	0.027	0.046	0.030	0.010	0.061	0.052	0.019	–0.031	–0.043	–0.018
C13	0.144	0.131	0.069	0.049	0.048	0.050	0.097	0.090	0.060	–0.095	–0.083	–0.019
C14	–0.027	–0.028	–0.025	0.058	0.068	0.084	0.016	0.020	0.029	0.086	0.097	0.109
O14′	0.063	0.058	0.044	0.127	0.129	0.129	0.095	0.093	0.086	0.064	0.071	0.085
C15	–0.018	–0.017	–0.009	–0.011	–0.010	–0.010	–0.014	–0.014	–0.010	0.007	0.006	–0.001
C16	–0.003	–0.002	–0.003	–0.006	–0.006	–0.006	–0.005	–0.004	–0.004	–0.003	–0.004	–0.003
C17	–0.002	–0.002	–0.003	–0.006	–0.006	–0.006	–0.004	–0.004	–0.004	–0.004	–0.004	–0.003
C18	–0.006	–0.006	–0.006	–0.006	–0.006	–0.005	–0.006	–0.006	–0.006	0.000	0.000	0.001
C19	–0.007	–0.007	–0.007	–0.007	–0.007	–0.006	–0.007	–0.007	–0.007	0.000	0.001	0.001
C20	0.006	0.007	0.009	–0.006	–0.005	–0.005	0.000	0.001	0.002	–0.012	–0.012	–0.014
R	0.019	–0.013	0.053	0.014	–0.012	0.039	0.017	–0.013	0.046	–0.005	0.002	–0.014

The O-substitution of the hydroxyl at position 12 has a noticeable impact on the Fukui functions f_k^- values for electrophilic attack at C-6, C-9, and C-13. While higher propensity for an electrophilic addition in 1 is observed at C-13 (with comparable descriptor at C-6 and C-9), subtle changes are observed in 6, and higher reactivity should be expected at C-6 for 7, while position C-13 is less prone for such reaction. Although tosylation of O-12′ has a significant impact on f_k^- when compared with 1 (probably due to its electron-withdrawing nature), O-methylation (electron-donating group) in 6 does not change the value significantly.

Examination of the Fukui functions f_k⁺ suggests that the oxygen atoms of the quinone system are likely to have the highest propensity toward a nucleophilic attack, regardless the O-substitution, which is in line with its propensity for reduction. A noticeable decrease of this index was determined for C-6 and C-11 upon tosylation in 7, while only subtle differences for such positions were visible when comparing 1 and 6. Indeed, the propensity for a nucleophilic attack in 7 at C-11 and C-14 is similar, contrasting with the higher electrophilic character of C-11 compared with C-14 in 1 and 6. Considering the susceptibility for reacting with a radical, while O-methylation seems to have little effect on radical susceptibility of C-9, O-tosylation in 7 considerably increases the reactivity at such position. This is also enforced by the electron’s withdrawal of tosyl group and the donation of electrons by the methyl group.

Further information can be obtained from the analysis of the dual descriptor. 1 and 6 are more likely to react as electrophiles upon reaction at C-11 with a nucleophile, while C-7 will be the preferred site for electrophilic reaction. Although a similar trend is visible for 7, the value of the dual descriptors is significantly altered in comparison with 1 and 6. From Table 1, is visible that C-7 is more likely to react with electrophiles in 7 than in 1 and 6, while an opposing tendency is observed for C-11, that is, this carbon is less likely to react with nucleophiles in 7 when compared with 1 and 6. Relevantly, introduction of O-tosylate changes the electronic nature at C-12, with this position being somewhat more likely to undergo nucleophilic attack in 7, rather than electrophilic reaction as expected for 1 and 6.

The electron density analysis of electrostatic potential surfaces of the three compounds (Figure 2) shows a localized electron density around the tosyl group in 7, while a deficit of density is observed in the same region in 1 due to the acidity of the enol. Overall, this analysis might suggest that the higher decomposition of the tosylated derivative 7 compared to 1 and 6 can come from two main factors: the increased reactivity toward radicals at position C-9 and increased reactivity of C-7 toward electrophiles.

Figure 2.

Total electron density mapped with the molecular electrostatic potential surfaces of 1, 6, and 7.

The above analysis on local reactivity descriptors was important not only to understand possible degradation pathways but also to guide further synthetic approaches toward new royleanone derivatives. To that matter, studying the compounds that were successfully synthesized and that efficiently inhibited P-gp (4 and 5) is also of the utmost importance. Herein, both molecular docking (for compounds 1–5) and molecular dynamics studies (4 and 5) were performed to clarify the molecular mechanisms by which they exert their inhibitory effect. Additionally, three other theoretical derivatives, 7α-acetoxy-6β-(4-chloro)benzoyloxyroyleanone (4a), 7α-acetoxy-12-O-(4-chloro)benzoylroyleanone (4b), and 7α-acetoxy-6β-benzoyloxyroyleanone (5a) were studied to evaluate the importance of the substituents and its substitution pattern in each position suitable for derivatization. In both studies, a refined murine P-gp structure²² and the location of drug-binding sites (M- R- and H-sites) were retrieved from a previous publication (additional data on the drug-binding sites available at the Supporting Information).²³

The herein results (Table 2) clearly show that the presence of additional aromatic moieties increases compound 4 binding affinity toward P-gp, decreasing its ΔG_bind by 2.8–3.1 kcal mol^–1 and 0.9–1.8 kcal mol^–1 (for R and M sites, respectively). Interestingly, the presence of additional binding poses at the substrate-binding H site (−10.5 kcal mol^–1) suggests an additional competitive mechanism, occurring at both the substrate-binding R- and H-sites, as one of the principal modes of action for royleanone derivative 4. As similar results were obtained for compound 5 and 5a (Table 2), the docking results seem to suggest that moieties at position C-6 may render more potent inhibitors due to the additional interactions with the Hoescht-33342 binding H-site (additional information on the P-gp drug-binding sites provided in the Supporting Information).

Table 2. Docking Results, Cross-Interaction Capability (CIc), and Protein–Ligand Contacts for Tested Compounds Obtained by LigPlot^a.

	top-ranked affinity kcal mol–1 (nr. of poses)				residues (no. interactions)
	H-site	R-site	M-site	CIc	Aro	Hyd	Pol
1		–8.8 (4)	–8.6 (5)	M	5 (23)	2 (2)	4 (8)
2		–8.7 (4)	–9.2 (5)	W	4 (10)	5 (9)	4 (3)*
3		–8.5 (4)	–9.5 (5)	W	6 (24)	2 (4)	5 (10)
4	–10.5 (2)	–11.6 (3)	–10.4 (4)	S	5 (13)	5 (13)	3 (4)*
4a	–9.2 (1)	–10.2 (4)	–10.3 (4)	S	5 (14)	4 (9)	2 (8)**
4b		–9.3 (3)	–9.2 (6)	M	5 (11)	4 (5)	4 (11)**
5		–9.3 (4)	–9.9 (5)	S	5(17)	4(9)	5(13)+
5a	–9.2 (1)	–10.4 (2)	–10.7 (6)	S	5(15)	5(7)	3(3)*

^aBinding energies refer to the top-ranked pose at each of the considered drug-binding sites. Hydrogen bonds with Ser697 and Ser943 (∗), with Ser693 (∗∗), or Thr297 (+) were observed. From a previous study,²³ docking binding energies for Rhodamine-123 were −9.2 kcal mol^–1 at both M- and R-sites. Regarding CIc: W, weak; M, moderate; S, strong (more information on CIc in the Supporting Information).

Nonetheless, low binding energies together with a strong cross-interaction capability (CIc)²³ when at the M-site also allow us to infer that 4 and 5 may also act as a noncompetitive efflux inhibitors. From the additional scaffolds (compounds 4a, 4b, and 5a), only the substituent at position C-6 seems particularly relevant for improving the biological activity, encompassing good binding affinities at both M- and R-sites, strong CIc values at the M-site and additional interactions with the H-site. Interestingly, when at the position C-12, the presence of the additional chlorine at the para-position of the substituent (i) seems to have a detrimental effect on the CIc (from strong to moderate) but without significantly changing the binding energies and (ii) is enough to shift the top-ranked binding pose of compound 4 from the M- to the substrate-binding R-site.

Docking results also confirm that for compound 4 (Figure S1B), while the 4-chlorobenzoyl substituent at position C-6 is deeply buried at the top of the M-site, in close vicinity to Phe71, Thr75, and Leu971, the substituent at position C-12 is oriented toward the water-filled cavity and not participating in significant interactions with the protein. Similarly, in all monosubstituted compounds, the substituent is also found to be buried and close to the apex of the M-site. Therefore, the initial data obtained from molecular docking studies uphold the importance of the benzoyloxy substituent for the inhibitory activity displayed by the royleanone derivatives while suggesting that in the other position this moiety can be replaced by other moieties either more prone to establish additional interactions with P-gp values or to fine-tune the physicochemical properties of the derivatives.

From the top-ranked poses obtained from the docking studies, a molecular dynamics (MD) study followed. Herein and as only 4(15) and 5 revealed to be stronger P-gp inhibitors, as assessed by a Rhodamine-123 accumulation assay, the MD approached focused on unveiling the molecular mechanism by which these compounds are able to inhibit efflux. To assess the importance of C-6/C-12 substituents in conferring stability to the derivatives under study, the relative importance of the substituents at both positions was additionally evaluated by studying the theoretical derivatives 4a, 4b, and 5a. The 40 ns MD simulations were performed. Herein, and due to the high flexibility of the transporter, longer MD simulations are expected to mask any effects that the ligand may have in the transporter and vice versa. Thus, only the last 20 ns of the MD run were considered production and analyzed.

First, the relative free-energy of binding was estimated through a MM/PBSA approach, corrected to encompass the presence of a lipid bilayer through an implicit membrane approach (Table 3).²⁴ From the data obtained, it is possible to infer that (i) the presence of a second benzoyloxy moiety does not significantly contribute to the binding affinity toward P-gp and that (ii) the presence of a para-substitution decreases the calculated ΔG_bind by ∼10 kcal mol^–1. These results agree with the docking predictions in which only one benzoyloxy moiety is needed to promote stronger interactions with P-gp, but nonetheless, they allow us to infer additional insights about the role that aromatic substitution may have on the interaction with the transporter. For comparison purposes, the relative free energy of binding of compound 4 at the substrate-binding R-site was calculated to be −71.0 ± 5.4 kcal mol^–1.

Table 3. Data from MD Simulations for Compounds 4, 4a, 4b, 5, and 5a^a.

	4	4a	4b	5	5a
Molecular Mechanics/Poisson–Boltzmann Surface Area
ΔGbind kcal mol–1 (±SD)	–78.5 (±4.4)	–72.4 (±4.4)	–70.0 (±4.0)	–87.7 (±5.0)	–84.2 (±4.6)
Hydrogen Bonding
<NHB>	1.806	1.861	1.413	1.731	0.861
ΔGHB (kJ mol–1)	–20.8	–24.8	–25.1	–21.7	–20.1
τ (ps)	490	2363	2613	700	377
Protein–Ligand Interactions
ligand efficiency	0.29 (7/23)	0.30 (6/20)	0.08 (2/23)	0.27 (6/22)	0.29 (5/17)
CIc	0.71	0.66	0.55	0.95	0.92
Oil–Water Partition Coefficient
molinspiration log P	8.91	6.72	6.38	5.45	5.79

^aΔG_bind, relative free-energy of binding; ΔG_HB, hydrogen-bond formation energy; <N_HB>, average number of hydrogen-bonds per time frame; τ, hydrogen bond lifetime; CIc, cross-interactions capability; logP, octanol–water partition coefficient.

The analysis of the protein–ligand hydrogen-bond network revealed that the existing HB network is reinforced through a decrease in HB formation energy while the HB lifetime increases (Table 3), due to interactions with Tyr303/Tyr949 (4a) and Gln191/Tyr949 (4b). For compounds 5 and 5a, HBs with Gln721 are the most predominant ones. Significant differences were also registered in ligand efficiency (LE), herein calculated as the ratio between the number of protein–ligand contacts with frequencies greater than 0.50 and the total number of registered protein–ligand contacts (for additional information see Methods section in the Supporting Information). Despite the HB reinforcement, compound 4b registered a dramatic decrease in its ligand efficiency (LE = 0.08) when compared with both compounds 4, 4a, 5, and 5a (LE = 0.29 ± 0.01). Herein, a closer inspection revealed that during the MD production run 4b shifted to a lower position inside the pocket (Figure S2), most probably due to the presence of the para-substituent. This was corroborated when evaluating CIc during the MD simulations. Compound 4 was calculated to be CIc = 0.71, slightly decreasing in 4a (CIc = 0.66) and having an even greater decrease in 4b (CIc = 0.55). Oppositely, both compounds 5 and 5a registered a substantial increase (CIc = 0.95 and 0.92, respectively). Therefore, it is clear that the presence of an aromatic para-substitution reduces the ability of the compounds in establishing sufficient cross-interactions to impair the conformational changes that are thought to mediate efflux, especially when the substituent is located at position C-12.

In addition to the results obtained from both molecular docking and molecular dynamics studies, the physicochemical properties of the compounds 4a, 4b, 5, and 5a, namely its octanol–water partition coefficient (log P),²⁵ become important because an effective P-gp inhibitor must be able to diffuse into and build up inside the lipid bilayer from the extracellular compartment.²⁶ Herein, the calculated log P for all monosubstituted derivatives is within the optimal range calculated for other diterpenic compounds, oppositely to that calculated for compound 4 (Table 3).

Experimental Validation

We have additionally confirmed our in silico findings by analyzing Rhodamine 123 accumulation in MDR cancer cells (Table 4). By means of compound 5 (the only one synthesized so far, having only one benzoyloxy substituent at position C-12, as confirmed by the crystal structure determined from single crystal X-ray diffraction data, see details in the Supporting Information), the observed MDR-reversal activity is similar to than in compound 4 and comparable to the control (Dex-VER), thus corroborating the importance of an unsubstituted aromatic substituent for the MDR reversal activity of all royleanone derivatives at position C-12. Further efforts are being undertaken to synthesize and evaluate the remaining compounds.

Table 4. P-gp Inhibition in an MDR Cancer Cell Line.

treatments	MFIa	FARb	SIc
NCI-H460 controld	134.10
NCI-H460/R controld	16.96
Dex-VERe	36.07	2.13	26.90
4	28.97	1.71	21.60
5	35.54	2.10	26.50

^aMeasured mean fluorescence intensity of Rhodamine 123 (MFI) was used for the calculation of the fluorescence activity ratio (FAR).

^bVia the following equation: FAR = MFI_MDR treated/MFI_MDR control. FAR values above 1.50 indicate P-gp inhibition.

^cSensitivity index (SI) was calculated on the basis of the measured mean fluorescence intensity (MFI) expressed via the following equation: SI = (MFI_MDR treated × 100)/MFI _{sensitive control}. SI values above 20 account for P-gp inhibition.

^dSensitive cancer cell line and its MDR counterpart used in the study: nonsmall cell lung carcinoma-NSCLC (NCI-H460 and NCI-H460/R).

^eDex-VER was applied at the same concentration (20 μM) as tested compounds.

Conclusions

The herein in silico studies provided valuable information to guide the design of future royleanone derivatives. While the analysis on local reactivity descriptors allowed us to understand possible degradation pathways and to guide further synthetic approaches toward new royleanone derivatives, molecular docking and MD studies further suggested that steric factors also play a role when choosing which substituent can be placed at position C-12. Furthermore, while the interaction with the substrate-binding H site suggests an additional competitive mechanism as the principal mode of action for RoyBzCl (4), favorable binding energies together with very high cross-interaction capabilities at the M-site (strong CIc) also allowed us to infer that compounds 4 and 5 (and its theoretical derivatives 4a, 4b, and 5a) may act as a noncompetitive efflux modulators when bound to the M-site.

Therefore, a suitable approach for the future generation of novel derivatives will involve the selective modification of position C-12 with alternative substituents (e.g., unhindered 5- to 6-member heterocycles, different linkers as amides, or conjugation with amino acids as glycine or glutamate) while varying the substitution pattern of the benzoyloxy moiety at position C-6.

Glossary

Abbreviations

AHR

7α-acetoxy-6β-hydroxyroyleanone

Clc

cross-interaction capability

Dex-VER

Dex-verapamil

DHR

6,7-dehydroroyleanone

FAR

fluorescence activity ratio

Δf_k

dual descriptor

ΔG_bind

relative free-energy of binding

HB

hydrogen bond; LE, ligand efficiency

logP

octanol–water partition coefficient

MD

molecular dynamics

MDR

multidrug resistance

MFI

fluorescent intencity of Rhodamine-123

P-gp

P-glycoprotein

SI

sensitivity index.

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications Web site. The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.9b00642.

• Computational details (methodology) on molecular descriptors calculations, molecular docking and molecular dynamics; details on crystal structure of 5; crystallographic information files were deposited at Cambridge Structural Database (CCDC 1986085) (PDF)

Author Contributions

^○ Shared senior authorship.

Support for this work was provided by FCT through UID/DTP/04567/2019, PTDC/MED-QUI/30591/2017, UIDB/00100/2020, and UIDP/00100/2020 and PhD grant SFRH/BD/137671/2018. The Academy of Finland is acknowledged for the financial support to N.R.C. (Decisions No. 326486/7) and Finnish Cultural Foundation for the financial support to S.H. (190336).

The authors declare no competing financial interest.