Labdane-Type Diterpenes from the Aerial Parts of Rydingia persica: Their Absolute Configurations and Protective Effects on LPS-Induced Inflammation in Keratinocytes

Abstract

Phytochemical investigations of an extract of the aerial parts of Rydingia persica led to the isolation of 14 labdane-type diterpenoids, of which compounds 1–5, 8, and 12–14 turned out to be new natural products, while the remaining compounds were isolated for the first time from the genus Rydingia. Their structures were elucidated using 1D- and 2D-NMR and mass spectrometry, and their absolute configurations were determined by quantum chemical calculation methods. Furthermore, DP4+ NMR chemical shift probability calculations were performed for compounds 12–14, in order to elucidate the orientation of the ambiguous chiral center at C-15, prior to absolute configuration determination. The methanol extract of the aerial parts of R. persica along with subfractions obtained and selected isolated compounds were evaluated for their effects on inflammation-related factors such as nitrotyrosine formation, IL-6 release, and TNF-α release, along with tight-junction proteins claudin-1 and occludin expression in LPS-stimulated HaCaT cells. Occludin and claudin-1 are tight-junction proteins, which play a pivotal role in wound repair mechanisms. Overall, the subfractions and compounds isolated showed moderate to high activity, indicating that labdane-type diterpenoids contribute to the anti-inflammatory and wound-healing activity of R. persica.

Inflammation is a hallmark of the initial response of the body to stimuli such as bacterial infection, tissue injury, and external irritants. In tissue injury, the innate system activation triggers a local inflammatory response, including the recruitment of inflammatory cells from the circulation, with the aim of removing the stimuli and to initiate local tissue recovery.¹ Inflammation occurs in almost every tissue but is maybe more obvious when the skin is affected, which up to 95% consists of keratinocytes forming the human epidermis. Therefore, inflammation of keratinocytes is associated with many skin problems.² Although inflammation is a prerequisite for healing, in the case of wounds and in general for skin restoration, uncontrolled inflammation can lead to tissue/skin damage and injuries. Consequently, alleviation of inflammatory injury is a pivotal step in the prevention of skin/tissue injury progression.³ The inflammatory phase intersects remarkably with primary hemostasis, occurring during the first 72 h after tissue/skin injury.⁴ This stage is associated with a complex series of molecular signals and eventually facilitates neutrophil and monocyte infiltration to the wound area, resulting in restraint of further tissue damage and elimination of pathogenic organisms and external remnants.⁴ TNF-α, IL-6, and IL-8 are some of the most important pro-inflammatory factors, which can be produced by a variety of cell types such as keratinocytes, macrophages, and mast cells.^1,5 In an early phase of inflammation, TNF-α attracts neutrophils and macrophages to the wounded area through chemokines.⁶ TNF-α is also involved in the repair of injured skin by inducing angiogenesis, resulting in an alleviated wound-healing process.⁶ IL-6 is a pleiotropic cytokine with an important role in the growth and differentiation of numerous cell types. In the skin, it is produced primarily by keratocytes and is detectable in cutaneous wounds.⁷ IL-8, as another pleiotropic inflammatory cytokine, plays a critical role in migration of immune system cells (mainly neutrophils) and leads to an increase in the expression of adhesion molecules in endothelial cells.¹ Overexpression of IL-6 and IL-8 in the wound area leads to prolonged inflammation and, therefore, to a delay in healing processes of injured skin/tissue. Previous studies have indicated the role of reactive oxygen species (ROS) as regulatory signals involved in wound-healing processes. Although low concentrations of ROS are required for maintenance of cell hemostasis and defense against invading microorganisms, excessive production results in oxidative damage and nonhealing wounds.⁸ Moreover, ROS are also mediators in inflammatory processes, thus promoting inflammation.⁹ For the barrier function, and restitution in the case of damage, tight-junction proteins such as occludin and claudin-1 are considered to be involved in proliferation and differentiation of keratinocytes.¹⁰ However, in the course of chronic inflammation, the expression of these two cell adhesion proteins is remarkably reduced, resulting in retardation of wound-healing processes.¹⁰ In healthy individuals, the wound-healing processes usually take place in a felicitous time period, leading to wound resolution. A delayed healing process could possibly arise from consecutive inflammation responses, infections, or insufficient angiogenesis responses.¹¹ Furthermore, elderly individuals and people suffering from diseases such as obesity and diabetes might undergo nonhealing chronic wounds, arising from prolonged inflammatory responses, consistent oxidative stress, and delayed re-epithelization, which remains as one of the major public health problems to be tackled. Over the past few decades, the search for natural product leads for the development of remedies for acceleration and improvement of healing processes has dramatically increased.¹² In this regard, ethnopharmacological studies provide valuable hints for the identification of plant constituents with potential wound-healing activities, since plants are often rich in antioxidants and display a high anti-inflammatory activity.

The genus Rydingia Scheen & V.A.Albert, belonging to the Lamiaceae family, encompasses currently four species worldwide. It was generated during the taxonomic revision of the genus Otostegia, which was reorganized to Otostegia sensu strictu and Rydingia and the transfer of some species to the genera Isoleucas and Moluccella.¹³ The four species of the genus Rydingia (R. integrifolia (Benth.) Scheen & V.A.Albert, R. limbata (Benth.) Scheen & V.A.Albert, R. michauxii (Briq.) Scheen & V.A.Albert, and R. persica (Burm.f.) Scheen & V.A.Albert) share some distinct morphological characters like spines at the leaf axils as well as spinose and persistent bracteoles, yellow flowers, and few-flowered verticillaster.¹³R. integrifolia is endemic to Eritrea, Ethiopia, and Yemen, while the other three species are endemic to Pakistan, Iran, and/or the Western Himalayas. Specifically, R. persica (syn. Otostegia persica (Burm.f.) Boiss.) is distributed mainly in the arid and subtropical regions of Iran, particularly in the southern and southeastern provinces.¹⁴ The aerial parts of R. persica were traditionally used by local tribes of the region and in Iranian traditional medicine as an insect repellent and for the treatment of diabetes, rheumatism, cardiac diseases, gastric discomfort, and wounds.^14,15 A previous in vivo study of a methanolic extract of R. persica revealed that the extract could significantly improve the wound-healing process.¹⁵ However, the study was not supported by either an appropriate phytochemical characterization or the identification of the active principles of R. persica responsible for facilitating the wound-healing process.

In view of these facts, this study aimed at the isolation and identification of pharmacologically relevant constituents of the aerial parts of R. persica. The absolute configuration of the isolated compounds was determined by electronic circular dichroism (ECD) and DP4+ NMR chemical shift probability calculation, in the case of ambiguous chiral centers. Recent years have seen an increasing number of applications of the DP4+ approach for the stereochemical assignment of challenging molecules. DP4+ probability-based chemical shift analysis utilizes geometries optimized in a higher level of theory than similar approaches such as DP4, and by adding the unscaled shift values the performance of the method increases significantly, providing more accurate and confident results in deciphering the relative configuration of challenging molecules.¹⁶ The effects of the isolated compounds on inflammation-related factors such as nitrotyrosine formation, IL-6 and TNF-α release, and claudin-1 and occludin expression were also studied in lipopolysaccharide (LPS)-stimulated HaCaT cells.

Results and Discussion

Aerial parts of R. persica were extracted with methanol, and the extract obtained was subjected subsequently to liquid–liquid extraction using solvents of different polarity (for details see the Experimental Section). Subfractions along with the MeOH extract showed pronounced inhibitory activities on IL-6 and TNF-α levels as well as nitrotyrosine formation of LPS-triggered HaCaT cells. Furthermore, they attenuated LPS-induced downregulation of claudin-1 and occludin (Tables S8 and S9, Supporting Information). None of these extract/subfractions showed any significant effect on cell viability in the concentration range 10–100 μg/mL (Table S6, Supporting Information). Based on these pharmacological results and the outcome of the LC-MS profiling of all subfractions, the petroleum ether subfraction was selected for further phytochemical investigations. Altogether, 14 labdane-type diterpenoids could be isolated (1–14; Figure 1). Among these, compounds 1–5, 8, and 12–14 were identified as new natural products. Although compounds 6,¹⁷7,¹⁸ leojapone A (9),¹⁹10,²⁰ and leosibirone A (11)^21,22 (Tables S1 and S2, Supporting Information) have been isolated from other plant species before, they were isolated and reported in the present investigation for the first time from the genus Rydingia.

Figure 1.

Structures of compounds isolated (1–14) from the aerial parts of R. persica.

Compound 1 (persicunin A) was isolated as a solid amorphous gum. The HRESIMS showed a [M + H]⁺ ion at m/z 375.2169 (calcd for C₂₂H₃₁O₅ [M + H]⁺, m/z 375.2166), corresponding to a molecular formula of C₂₂H₃₀O₅. The UV spectrum (CH₃CN) displayed maxima at 208 (2.37) and 247 (2.54) nm. Analysis of the ¹H NMR along with the HSQC spectra indicated resonances of four methyl groups at δ_H 1.80 (H-17), 1.11 (H-20), 1.08 (H-18), and 2.07 (H-22); six methylene groups at δ_H 2.58 (H-11), 2.47 (H-12), 2.00 (H-2a), 1.78 (H-2b), 2.00 (H-1a), 1.74 (H-1b), 2.47 (H-6), 4.25 (H-19a), and 4.02 (H-19b); and two aliphatic methine groups at δ_H 3.76 (H-3) and 2.36 (H-5). Furthermore, analysis of the ¹³C NMR spectrum revealed 22 carbons, of which two carbonyl groups at δ_C 199.4 (C-7) and 171.2 (C-21) and six olefinic carbons at δ_C 130.7 (C-8), 166.3 (C-9), 124.5 (C-13), 110.7 (C-14), 143.2 (C-15), and 138.8 (C-16) were characteristic. Subsequently, the COSY spectrum depicted two spin systems of H-1/H-2/H-3 and H-15/H-14/H-16 (Figure 2). The HMBC spectrum displayed correlations from H-1 to C-5, C-9, C-20 and C-3 and from H-3 to C-4, C-5, C-18, and the methylene group C-19. Considering the further correlations from H-5 to the carbonyl group at δ_C 199.4 (C-7) and olefinic carbons at δ_C 166.3 (C-9), 130.7 (C-9), and 166.3 (C-8), a modified decalin unit was readily assigned.

Figure 2.

Key HMBC and COSY correlations of new compounds 1–5, 8, and 12–14.

Additionally, HMBC correlations from H-5 and H-18 to C-19 and from H-19a,b and H-22 to the carboxyl group at δ_C 171.2 (C-21) led to assignment of an acetyl group at position C-19. Analysis of the HMBC correlations of a second spin system obtained from the COSY spectrum indicated correlations from H-15 to C-16, C-14, and C-13; from H-14 to C-16, C-13, and C-12; and from H-16 to C-13 and C-12. Considering the downfield resonances of C-15 and C-16, an oxygen bridge was established to connect these two atoms, resulting in the elucidation of a furan ring at position C-12. Furthermore, it was found that the decalin and furan units are connected through two adjacent methylene groups (C-11 and C-12), which was confirmed by correlations from H-12 to C-16 and C-9 and from H-11 to C-13. Therefore, compound 1 was established as (9-(12-(furan-13-yl)ethyl)-3-hydroxy-4,17,20-trimethyl-7-oxo-1,2,3,4,5,6,7, 10-octahydronaphthalen-4-yl)methyl acetate (Figures 1,2), assignable as a furano-labdane-type diterpenoid.

The NOESY spectrum revealed correlations from H-20 to H-19a and H-19b as well as from H-19a to H-3, which together with the lack of a correlation from H-20 to H-5 led to the determination of the relative configuration of 1 as 3R*, 4S*, 5R*, and 10S* (Figure 3). ECD simulation of compound 1 at the TD-DFT/B3LYP/6-31G(d,p)//B3LYP/6-31G(d,p) level in acetonitrile and a comparison with the experimental ECD spectrum (Figure 4) resulted in the determination of the absolute configuration of 1 as 3R, 4S, 5R, and 10S, and this compound was named persicunin A.

Figure 3.

Key NOESY correlations of isolated new compounds 1–5, 8, and 12–14 (for R, refer to Figure 1).

Figure 4.

Comparison of the experimental and calculated ECD spectra of compound 1 and the experimental ECD spectrum of compound 2.

Compound 2 (persicunin B) was isolated as a solid amorphous gum. The HRESIMS showed a [M + H]⁺ ion at m/z 407.2059 (calcd for C₂₂H₃₁O₇ [M + H]⁺, m/z 407.2064), corresponding to a molecular formula of C₂₂H₃₀O₇. The UV spectrum (CH₃CN) displayed maxima at 200 (2.95) and 246 (2.75) nm. Analysis of the ¹H NMR and HSQC spectra indicated similar chemical shifts to those of compound 1 (Table 1), supporting the presence of a labdane-type scaffold for this compound. Detailed investigation of the ¹³C NMR and HMBC spectra revealed differences in the furan ring in comparison to 1. In fact, H-11/12 exhibited HMBC correlations to a lactone carboxyl group at δ_C 171.2 (C-16), a quaternary olefinic carbon at δ_C 137.6 (C-13), and an olefinic methine group at δ_C 144.1 (C-14). Furthermore, H-14 showed a correlation to an aliphatic methine carbon at δ_C 97.8 (C-15), the proton of which (H-15) revealed a correlation to a carboxyl group at δ_C 171.2 (C-16). Considering the downfield resonance of H-15 (δ_H 7.37) and the chemical shift of the carboxyl group C-16 (δ_C 171.3), the furan ring of 1 had to be replaced by a lactone ring in compound 2 with a hydroxy group at position C-15. Therefore, the structure of 2 was deduced as (3-hydroxy-9-(12-(15-hydroxy-16-oxo-15,16-dihydrofuran-13-yl)ethyl)-4,17,20-trimethyl-7-oxo-1,2,3,4,5,6,7,10-octahydronaphthalen-4-yl)methyl acetate (Figure 1). The NOESY spectrum revealed similar NOEs to those of 1 (Figure 3), leading to the determination of an identical relative configuration for the decalin unit (3R*, 4S*, 5R*, and 10S*). However, the relative configuration of C-15 could not be determined due to (i) possible epimerization of this center resulting from its hemiacetalic character and (ii) the lack of NOEs between this unit and the decalin unit. Finally, comparison of the ECD spectra of 1 and 2 enabled the determination of the absolute configuration of 2 as 3R, 4S, 5R, and 10S (Figure 4).

Table 1. ¹H NMR Spectroscopic Data (δ_H in ppm, J in Hz) of Compounds 1–5, 8, and 12–14 in CHCl₃-d or MeOH-d₄ (600.19 MHz).

position	1	2	3	4	5	8	12	13a	14
1	2.00, m	1.85, d (11.6)	1.93, m	1.65, dt (12.5, 2.16)	1.70, dq (11.8, 4.0, 3.4)	2.01, m	2.56, td (14.1, 4.3)	1.96, m	1.94, m
	1.74, m	1.71, dt (12.7, 3.5)	1.64, m		1.93, m	1.35, m	1.16, m	1.24, dt (6.5, 3.0)	1.12, m
2	2.00, m	1.98, tt (14.3, 2.9)	2.04, m	2.01, tdd (13.77, 4.20, 2.56)	1.74, ddd (14.4, 3.5, 2.0)	1.73, dd (14.5, 4.6)	2.02, m	1.66, ddd (13.1, 6.3, 3.0)	2.04, ddd, (14.3, 3.9, 2.4)
	1.78, m	1.78, m	1.74, m	1.73, dt (14.59, 7.88, 2.88)	2.04, m		1.77, ddd (14.6, 8.2, 3.5)	2.02, m	1.66, ddd (14.2, 7.6, 3.5)
3	3.76, br s	3.76, d (2.8)	3.53, t (2.4)	3.52, pseudo t (2.9)	3.50, t (2.8)	3.46, br s	3.41, m	3.40, t (2.5)	3.33, m
5	2.36, dd (14.4, 3.5)	2.30, dd (14.4, 3.9)	2.22, dd (7.0, 11.2)	2.17, m	2.22, ddd (12.2, 6.2, 3.0)	2.49, m	2.43, dd (14.1, 3.8)	2.40, d	2.17, m
6	2.47, mb	2.44, dd (17.3, 14.4)	2.40, m	2.41, m	2.40, d (7.5)	2.32, d (4.5)	2.38, m	2.40, m	4.24, pseudo t (3.4)
			2.43 m				2.24, m	2.22, m
8						2.72, d (6.7)	2.72, d (6.6)	2.82, q (6.6)	3.42, d (6.7)
11	2.58, m	2.50, m	2.47, m	2.51, m	2.45, mb	1.99, m	2.30, m	2.27, m	2.22, m
			2.56, m			1.86, ddd (14.5, 11.4, 5.6)	2.04, m	1.97, m	1.90, m
12	2.47, mb	2.50, m	2.47, m	2.51, m	2.49, mb	2.54, ddd (14.2, 11.6, 5.5)	2.37, m	2.08, m	2.11, m
			2.56, m			2.45, m	2.27, m	1.98, m	1.90, m
14	6.30, m	6.87, s	5.92, s	6.86, br s	6.84, t (1.3)	6.27, br s	2.49, dd (13.3, 5.9)	2.21, m	2.25, m
							2.19, d (13.4, 0.7)	2.21, m	2.08, m
15	7.37, t (1.6)	6.13, s		6.09, br s	5.75, dd (2.7, 1.3)	7.36, pseudo t (1.7)	5.36, dd, (5.9, 0.7)	4.94, dd (5.6, 4.1)	4.94, dd (5.7, 3.3)
16	7.26, m		6.04, s			7.23, br s		3.74, d (8.1)	3.88, d (8.3)
								3.54, d (8.1)	3.58, d (8.3)
17	1.80, s	1.81, s	1.77, s	1.81, s	1.79, s	1.13, d (6.7)	1.03, d (6.6)	0.95, d (6.6)	1.01, d (6.7)
18	1.08, s	1.07, s	0.97, s	0.96, s	0.97, s	0.96, s	0.97, s	0.92, s	1.28, s
19	4.25, d (11.4)	4.22, d (11.4)	0.96, s	0.94, s	0.95, s	0.93, s	0.89, s	0.91, s	1.07, s
	4.02, d (11.4)	4.01, d (11.5)
20	1.11, s	1.09, s	1.12, s	1.08, s	1.10, s	1.19, s	1.17, s	1.19, s	1.43, s
21					3.59, s		3.47, s	5.32, s	3.33, s
22	2.07, s	2.07, s

^aMeOH-d₄ used as NMR solvent.

^bInterchangeable signals.

Compound 3 (persicunin C) was isolated as a pale yellow, amorphous gum. Its molecular formula was deduced as C₂₀H₂₈O₅ based on HRESIMS data (found m/z 349.2008, calcd for C₂₀H₂₉O₅ [M + H]⁺, m/z 349.2010). The UV spectrum (CH₃CN) displayed an absorption maximum at 247 (2.13) nm. The ¹H NMR and HSQC spectra indicated the presence of four methyl groups at δ_H 1.77 (H-17), 1.12 (H-20), 0.97 (H-18), and 0.96 (H-19); five methylene groups at δ_H 2.54 (H-11a, H-12a), 2.47 (H-11b, H-12b), 2.04 (H-2a), 1.74 (H-2b), 1.93 (H-1a), and 1.64 (H-1b); and three aliphatic methine groups at δ_H 6.04 (H-16), 3.53 (H-3), and 2.22 (H-5), along with the resonance of an olefinic methine proton at δ_H 5.92 (H-14). The ¹³C NMR spectrum depicted 20 carbon signals, of which four quaternary carbons at δ_C 168.3 (C-13), 165.3 (C-9), 131.2 (C-8), and 40.8 (C-10); a aliphatic methine carbon at δ_C 98.9 (C-16); and two carbonyl groups at δ_C 200.1 (C-7) and 170.8 (C-15) were characteristic. Analysis of the COSY spectrum revealed two spin systems of H-1/H-2/H-3 and H-5/H-6. Subsequent analysis of the HMBC spectrum indicated correlations from H-1b (δ_H 1.64) to C-10 (δ_C 40.8), C-5 (δ_C 43.6), C-3 (δ_C 75.1), and C-9 (δ_C 165.3); from H-3 (δ_H 3.53) to C-1 (δ_C 28.7), C4 (δ_C 37.6), C-5 (δ_C 43.6), and two singlet methyl groups at C-18 (δ_C 27.5) and C-19 (δ_C 21.8); and from H-5 (δ_H 2.22) to C-1 (δ_C 28.7), C-10 (δ_C 40.8), C-6 (δ_C 43.9), and C-9 (δ_C 165.3). Thus, again the presence of a decalin unit could be assumed. The presence and position of an α,β-unsaturated carbonyl group was confirmed by the following HMBC correlations: (i) H-6 (δ_H 2.41) to a ketone-type carbonyl group at δ_C 200.1 and a quaternary carbon of an olefinic bond at δ_C 131.2; (ii) H-20 (δ_H 1.12) and H-1a (δ_H 1.93) to C-9 (δ_C 165.3), and (iii) a methyl group at δ_H 1.77 (H-17) to C-8 (δ_C 131.2), C-9 (δ_C 165.3), and C-7 (δ_C 200.1). The protons of two adjacent methylene groups, H-11/12 (δ_H 2.51), displayed HMBC correlations with C-10 (δ_C 40.8), C-16 (δ_C 98.9), C-14 (δ_C 117.9), C-8 (δ_C 131.2), C-9 (δ_C 165.3), and C-13 (δ_C 168.7). Considering a further ³J correlation from H-16 (δ_H 6.04) to a carbonyl group at 170.8 (C-15), the presence of a lactone ring was suggested as being connected via an ethylene group to C-9 of the decalin moiety. Thus, the structure of this labdane-type diterpenoid was deduced as 16-hydroxy-13-(11-(3-hydroxy-4,4,17,20-tetramethyl-7-oxo-3,4,4a,5,6,7,8,8a-octahydronaphthalen-9-yl)ethyl)furan-15(16H)-one (Figure 1). Analysis of the NOESY spectrum revealed no correlation between H-20 and H-5, indicating the presence of a trans-decalin unit. Moreover, the NOE from H-5 to H-18 and the lack of correlation from H-5 to H-3 revealed an α-orientation of the hydroxy group at position C-3 (Figure 3). Therefore, the relative configuration of the decalin unit was determined as 3R*, 5R*, and 10S*. The relative configuration of C-16 could not be established due to both the lack of NOE interactions with other chiral centers and the racemization of C-16 due to its hemiacetal character. Since the ECD spectrum of 3 was similar to that of 7 (Figure 5), ECD simulation was performed for 7, and the results were further extended to 3 (and later for 4 and 5). Therefore, the absolute configuration of the decalin unit of 3 was determined as 3R, 5R, and 10S (Figure 1).

Figure 5.

Comparison of the experimental and calculated spectra of compound 7 (A) and overlaying of the experimental ECD spectra of 3–5 (B).

Compound 4 (persicunin D) was isolated as a whitish, amorphous gum. Its molecular formula was deduced as C₂₀H₂₈O₅ based on HRESIMS data (found m/z 347.1872, calcd for C₂₀H₂₇O₅ [M – H]⁻, m/z 347.1864). The UV spectrum (CH₃CN) displayed absorption maxima at 200 (3.51) and 247 (3.47) nm. The ¹H and ¹³C NMR spectra of 4 were closely comparable to those of 3 (Table 1), revealing again a labdane-type scaffold for this compound. Subsequent analysis of the HMBC spectrum revealed correlations from H-11/12 to a quaternary olefinic carbon at δ_C 137.6 (C-13), an olefinic methine carbon at δ_C 144.0 (C-14), and a lactone carboxyl at δ_C 171.3 (C-16). It was concluded that the position of the carbonyl group and methine group of the five-membered ring unit in 4 interchanged in comparison to 3. Therefore, the structure of 4 was deduced as 15-hydroxy-13-(11-(3-hydroxy-4,4,17,20 tetramethyl-7-oxo-1,2,3,4,5,6,7,10-octahydronaphthalen-9-yl)ethyl)furan-16(15H)-one (Figure 1). Analysis of the NOESY spectrum revealed similar correlations to those found for 3 (Figure 3), which resulted in determination of the relative configuration of compound 4 as 3R*, 5R*, and 10S*. Similar to 3, the relative configuration of C-15 could not be assigned due to the epimerization and a lack of appropriate NOE correlations. However, as the ECD of 4 resembled closely those of 3 and 7, the absolute configuration of the decalin unit of 4 was proposed as 3R, 5R, and 10S (Figure 1).

Compound 5 (persicunin E) was isolated as a colorless gum. The UV spectrum (CH₃CN) was similar to that of 3 (247 nm). The ¹H and ¹³C NMR spectra of 5 were very similar to 4, revealing analogous NMR data (Table 1). Its molecular formula was deduced as C₂₁H₂₈O₅ based on the HRESIMS (found m/z 363.2176, calcd for C₂₁H₃₁O₅ [M + H]⁺, m/z 363.2166), 14 Da higher than that of compound 4, suggesting an additional methyl group substitution. Analysis of the HMBC spectrum showed correlations from the protons of the additional methyl group at δ_H 3.59 (H-21) to the methine carbon at δ_C 102.4 (C-15), confirming the position of the methyl group. The NOESY spectrum displayed similar correlations to those of compound 4 and revealed again no conclusive relative configuration for C-15 (Figure 3). Since the proton signal of one methyl group (H/C-21), as well as the carbon signals of C-13, C-14, and C-15 showed a slight doubling, the presence of an epimeric mixture, due to the racemization of C-15, was assumed for compound 5. To determine the absolute configuration of the remaining stereocenters, the ECD spectrum of 5 in acetonitrile was compared to that of 4 and 7 (Figure 5), and considering all observed NOE correlations, the structure of 5 was deduced as 13-(11-((3R,5R,10S)-3-hydroxy-4,4,17,20-tetramethyl-7-oxo-1,2,3,4,5,6,7,10-octahydronaphthalen-9-yl)ethyl)-15-methoxyfuran-16(15H)-one (Figure 1).

Compound 8 (calyonol) was isolated as a colorless gum. Its molecular formula was deduced as C₂₀H₃₀O₄ based on HRESIMS data (found m/z 357.2033, calcd for C₂₀H₃₀O₄Na [M + Na]⁺, m/z 357.2036). The UV spectrum of 8 exhibited an absorption maximum at 210 (3.49) nm. Further analysis of the ¹H, ¹³C, and HSQC NMR spectra revealed the resonances of five methyl groups at δ_H 1.19 (H-20), 1.13 (H-17), 0.96 (H-19), and 0.93 (H-18); a carbonyl group at δ_C 211.8 (C-7); and characteristic protons of a furan ring at δ_H 7.36 (H-15), 7.23 (H-16), and 6.27 (H-14), confirming the presence of a furano-labdane-type diterpenoid structure. Further analysis of the HMBC contacts indicated hydration of the olefinic bond (C-8–C-9) and introduction of a hydroxyl group at C-9 (δ_C 81.8). Therefore, the structure of 8 was deduced as 9-(12-(furan-13-yl)ethyl)-3,9-dihydroxy-17,18,19,20-tetramethyl octahydronaphthalen-7(6H)-one (Figure 1). Analysis of the NOESY spectrum revealed correlations between H-20, H-8, and H-11, indicating the α-orientation of the hydroxy group at C-9, in accordance with the biosynthetic route for this class of compounds. Considering the other observed NOESY correlations (Figure 3), the relative configuration of 8 was determined as 3R*, 5S*, 8S*, 9R*, and 10S*. Moreover, the ECD spectrum of 8 was simulated using the TD-DFT/cam-B3LYP/6-31G(d,p)/CPCM level in acetonitrile, resulting in determination of the absolute configuration of 8 as 3R, 5S, 8S, 9R, and 10S (Figure 6) by comparison with the experimental spectrum. This compound was found to be the deacetylated derivative of calyone, isolated from Roylea calycina,²³ and therefore was named calyonol.

Figure 6.

Comparison of the experimental and calculated ECD spectra of 8.

Compound 12 (persicunin F) was obtained as a whitish gum. Its molecular formula was deduced as C₂₁H₃₂O₆ based on the HRESIMS data (found m/z 365.1936, calcd for C₂₀H₂₉O₆ [M – CH₃]⁻, m/z 365.1970). The UV spectrum (CH₃CN) displayed absorptions at 208 (3.66), 248 (3.31), and 313 (2.86) nm. The ¹H NMR and HSQC spectra revealed resonances for five methyl groups at δ_H 3.47 (H-21), 1.17 (H-20), 1.03 (H-17), 0.97 (H-19), and 0.89 (H-18); six methylene groups at δ_H 5.36 and 2.49 (H-14a and H-14b), 2.38 and 2.24 (H-6a and H-6b), 2.37 and 2.27 (H-12a and H-12b), 2.56 and 1.19 (H-1a and H-1b), 2.30 and 2.04 (H-11a and H-11b), and 2.02 and 1.77 (H-2a and H-2b); and four methine groups at δ_H 3.41 (H-3), 2.43 (H-5), 2.72 (H-8), and 2.19 (H-15). Further analysis of the ¹³C NMR spectrum indicated the presence of 21 carbons, of which four quaternary carbons at δ_C 98.6 (C-9), 84.1 (C-13), 42.7 (C-10), and 38.1 (C-4), along with one downfield methine group at 102.0 (C-15), a carbonyl group at 210.8 (C-7), and a carboxyl group at 177.9 (C-14), were readily distinguishable. The COSY spectrum exhibited three spin systems of H-15/H-16, H-8/H-17, and H-1/H-2/H-3 (Figure 3). Analysis of the HMBC spectrum unveiled correlations from H-1 to H-10, H-9, and H-20 and from H-5 to H-20, H-18, and H-10. Additional correlations from two methyl groups (H-18, H-19) to C-5 and C-3 resulted in the identification of a furano-labdane-type diterpenoid scaffold. The HMBC spectrum revealed also correlations from H-17 (δ_H 1.03) to C-11 (δ_C 28.9), C-8 (δ_C 49.7), and C-9 (δ_C 98.6) and from H-6a,b (δ_H 2.24 and 2.38) to C-7 (δ_C 210.8) and C-10 (δ_C 42.7). Additionally, two methylene groups, H-11a,b (δ_H 2.04, 2.30) and H-12a,b (δ_H 2.27, 2.37), disclosed correlations to the aliphatic carbon of the decalin ring (C-9) at δ_C 98.6. Due to the downfield resonance of C-9 compared to compound 8, it was concluded that C-9 is not only connected to a heteroatom like oxygen but also involved in an additional ring closing, consistent with the downfield shift of this carbon. Moreover, H-11, H-12, and H-16a,b exhibited correlations to a spiro carbon (C-13) at δ_C 84.1 and a lactone carboxyl group (C-14) at δ_C 177.9. Further correlations of H-16 to C-15 (observed also as COSY spin system) and H-15 to C-14 (lactone carboxyl) revealed the structure of the spiro ring. Therefore, the structure of 12 was deduced as 3-hydroxy-15-methoxy-17,18,19,20-tetramethyldecahydro-8H,16H-dispiro[furan-13-furan-9-naphthalene]-7,16(H)-dione (Figure 1). The NOESY spectrum exhibited NOEs between H-5 and H-18; H-3 and H-19; H-8 and H-20; and H-17 and H-14a,b. By taking into account the lack of NOESY correlations from H-20 to H-5, and H-17 to either of H-11 or H-12, the relative configuration of 12 was assigned as 3R*, 5S*, 8S*, 9R*, 10S*, and 13S*. However, due to overlapping signals, the relative configuration of the methoxy group at C-15 could not be assigned.

To overcome this problem, the DP4+ NMR chemical shift calculation method was applied to solve the relative configuration of C-15 (Figure S89, Supporting Information), which resulted in the assignment of an S* configuration for the respective carbon. Subsequent simulation of the ECD spectrum of compound 12 and its comparison with the experimental spectrum in acetonitrile (Figure 7) led to determination of the absolute configuration of 12 as 3R, 5S, 8S, 9R, 10S, 13S, and 15S.

Figure 7.

Comparison of experimental and calculated ECD spectra of compound 12.

Compound 13 (persicunin G) was obtained as a whitish gum. Its molecular formula was deduced as C₂₁H₃₄O₅ based on its HRESIMS data (found m/z 389.2294, calcd for C₂₁H₃₄O₅Na [M + Na]⁺, m/z 389.2298). The UV spectrum (CH₃CN) displayed an absorption maximum at 200 (3.17) nm. The ¹H and ¹³C NMR spectra were in accordance with a labdane-type structure, with a similar scaffold to compound 12. Detailed analysis of the ¹³C and HSQC spectra revealed a lack of the carbonyl group at δ_C 177.9 (C-16) present in 12, but with the occurrence of an oxygenated methylene group at δ_C 76.2 (C-16). HMBC correlations from H-11 (δ_H 1.97, 2.27) and/or H-12 (δ_H 1.98, 2.08) to C-16 were used to establish the position of this group. The NOESY spectrum revealed NOEs similar to those of 12 for the decalin unit and additionally from H-17 to H-16, indicating an inversion of the spiro ring in comparison to that of 12 (Figure 3). Considering all of the information obtained, compound 13 was elucidated as (3R*,5S*,8S*,9R*,10S*,13R*)-3-hydroxy-15-methoxy-17,18,19,20-tetramethyldecahydro-16H,8H-dispiro[furan-13-furan-9-naphthalen]-7(6H)-one (Figure 1). In a comparable manner to 12, the use of NOESY correlations to assign the relative configuration of C-15 was not conclusive. Therefore, the relative configuration of C-15 was determined by DP4+ NMR chemical shift probability calculation, revealing again an S* configuration (Figure S90, Supporting Information). Subsequently, its ECD spectrum was simulated at the TD-DFT/cam-B3LYP/6-31G(d,p)/CPCM/methanol level and compared with the experimentally obtained ECD spectrum in methanol (Figure 8), which led to the determination of the absolute configuration of 13 as 3R, 5S, 8S, 9R, 10S, 13R, and 15S.

Figure 8.

Comparison of the experimental and calculated ECD spectra of compound 13.

Compound 14 (persicunin H) was obtained as a whitish gum. The UV spectrum of 14 (CH₃CN) displayed an absorption maximum at 247 (2.43) nm. Its molecular formula was deduced as C₂₁H₃₄O₆, based on the HRESIMS data (found m/z 405.2228, calcd for C₂₁H₃₄O₆Na [M + Na]⁺, m/z 405.2248), showing a mass gain of 16 Da compared to that of 13. Analysis of ¹H and ¹³C NMR spectra along with the 2D-NMR data obtained suggested the presence of a close derivative of compound 13. The ¹H NMR and HSQC spectra displayed an additional methine group (C-6) at δ_H 4.20, in which the downfield resonance of its proton and carbon suggested the introduction of a hydroxy group at C-6, as confirmed by its COSY and HMBC correlations with H-5 (δ_H 2.17). On the other hand, the spectra again showed evidence for a methoxy group, as present also in compounds 12 and 13. The relative configuration of C-6 was found to be S*, as deduced from the ³J coupling constant of its proton and the adjacent proton (J = 3.4 Hz). Moreover, this configuration was supported by the NOEs observed between H-6/H-5/CH₃-19. The relative configuration of the other stereocenters was determined from the observed NOEs (Figure 3). Therefore, the structure of compound 14 was deduced as (3R*,5S*,6S*,8S*,9R*,10S*,13R*)-3,6-hydroxy-15-methoxy-17,18,19,20-tetramethyldecahydro-16H,8H-dispiro[furan-13-furan-9-naphthalen]-7(6H)-one. Similar to 13, the relative configuration of C-15 could not be determined. Therefore, the relative configuration of C-15 was determined again by performing a DP4+ NMR chemical shift probability calculation, but was assigned as R* in contrast to 13 (Figure S91, Supporting Information). Moreover, its ECD spectrum was simulated at the TD-DFT/cam-B3LYP/6-31G(d,p)/CPCM/acetonitrile level and compared with the experimental ECD spectrum in methanol (Figure 9), which led to the determination of the absolute configuration of 14 as 3R, 5S, 6S, 8S, 9R, 10S, 13R, and 15R.

Figure 9.

Comparison of the experimental and calculated ECD spectra of compound 14.

Keratinocyte Protective Activity of the Isolated Compounds

Cytokine production during skin inflammation is of crucial importance for insult response regulation and homeostasis restoration.²⁴ Chronic wounds show, in particular, elevated cytokine levels, including those of TNF-α, IL-6, IL-8, and IL-1β, which disrupt the fibroblast and keratinocyte crosstalk, which is critical for wound healing.²⁵ Thus, the inhibition of cytokine production or signaling is one of the main properties for compounds with a pharmacological potential in wound healing. In order to investigate the impact on this mode of action, the compounds isolated herein were examined for their inhibitory effects on IL-6 and TNF-α release. However, before pursuing such experiments, the antiproliferative activity and stability of the test compounds were considered. The latter was necessary since a previous investigation on similar labdane-type diterpenoids from Leonorus japonicus revealed that these classes of compounds are prone to degradation in polar media, such as water and MeOH.²⁰ Therefore, all compounds isolated from R. persica were subjected to a stability test in cell culture medium with 1% (v/v) DMSO at 37 °C for 24 h, prior to the cell culture experiments. All compounds with a degradation of less than 20%, evaluated as reduction of peak area in HPLC-DAD analysis, were considered for further pharmacological investigation (data not shown). In summary, 10 compounds (1, 2, 4–6, 8, 9, and 12–14; Tables 3 and 4) were found to be stable under the selected assay conditions. All selected compounds were tested further for their antiproliferative activity on HaCaT cells using an MTT assay.³⁷ As shown in Table S7 (Supporting Information), none of the compounds displayed remarkable antiproliferative activity in the range of the concentrations applied (100–12.5 μM).

Table 3. Effect of Compounds 1, 2, 4–6, 8, 9, and 12–14 Isolated from the Aerial Parts of R. persica on Inhibition of Nitrotyrosine Formation, IL-6 Release, and TNF-α Release vs LPS, on the HaCaT Cells.

compound	conc. (μM)	mean ± SEM of % of inhibition of IL-6 release vs LPSa	mean ± SEM of % of inhibition of TNF-α release vs LPSa	mean ± SEM % of inhibition of nitrotyrosine formation vs LPSa
1	50	73.50 ± 4.50***	44.00 ± 5.00**	58.33 ± 3.18***
	25	67.00 ± 2.00***	25.00 ± 0.00*	44.00 ± 2.00***
	12.5	66.50 ± 3.50***	0.00 ± 0.00	0.00 ± 0.00
2	50	73.50 ± 2.50***	56.50 ± 1.50***	61.67 ± 3.38***
	25	49.50 ± 3.50***	47.50 ± 1.50**	43.00 ± 3.05***
	12.5	44.50 ± 0.50***	35.00 ± 0.00*	3.67 ± 0.67
4	50	76.00 ± 4.00***	73.50 ± 3.50***	58.67 ± 6.17***
	25	70.50 ± 1.50***	70.50 ± 3.50***	37.00 ± 1.00***
	12.5	56.00 ± 2.00***	68.00 ± 1.00***	30.00 ± 1.52***
5	50	74.50 ± 2.50***	72.00 ± 0.00***	59.33 ± 1.67***
	25	74.00 ± 5.00***	68.00 ± 1.00***	29.50 ± 1.50**
	12.5	65.50 ± 2.50***	52.00 ± 3.00***	0.50 ± 0.50
6	50	81.00 ± 1.00***	68.00 ± 1.00***	56.00 ± 2.00***
	25	69.00 ± 1.00***	68.00 ± 1.00***	40.00 ± 0.00***
	12.5	65.50 ± 0.50***	39.50 ± 12.50*	5.00 ± 2.00
8	50	69.50 ± 6.50***	65.00 ± 4.00***	46.33 ± 3.90***
	25	64.50 ± 8.50***	62.50 ± 1.50***	30.50 ± 2.72***
	12.5	48.00 ± 7.00***	37.00 ± 2.00*	6.33 ± 2.19
9	50	80.50 ± 1.50***	68.00 ± 4.00***	56.33 ± 0.88***
	25	77.00 ± 1.00***	62.00 ± 7.00***	42.67 ± 3.28***
	12.5	73.50 ± 8.50***	15.62 ± 6.50	9.00 ± 2.52
12	50	76.50 ± 2.50***	56.50 ± 1.50***	57.50 ± 6.50***
	25	75.50 ± 0.50***	50.50 ± 1.50***	40.00 ± 1.53***
	12.5	75.00 ± 1.00***	41.50 ± 10.50*	14.33 ± 5.88
13	50	76.50 ± 3.50***	68.00 ± 1.00 ***	39.33 ± 4.33***
	25	71.50 ± 2.50***	62.50 ± 1.50***	21.00 ± 2.08
	12.5	63.00 ± 1.00***	44.00 ± 5.00**	0.00 ± 0.00
14	50	78.00 ± 2.00***	65.00 ± 7.00***	38.00 ± 4.93***
	25	68.50 ± 0.50***	55.00 ± 6.00***	15.00 ± 0.50
	12.5	38.50 ± 3.50***	44.00 ± 2.00**	0.00 ± 0.57
indomethacin	10	88.50 ± 1.5***	77.00 ± 1.00***

^aData are expressed as mean ± SEM of the percentage of inhibition vs cells treated with LPS alone of n = 3; ***, **, and * denote respectively p < 0.001, p < 0.01, and p < 0.05 vs cells treated with LPS alone. Indomethacin (10 μM) was used as a positive control for IL-6 and TNF-α release experiments.

Table 4. Effect of Compounds 1, 2, 4–6, 8, 9, and 12–14 Isolated from the Aerial Parts of R. persica on Claudin-1 and Occludin Expression (Mean Fluorescence Intensity) in HaCaT Cells after Treatment with LPS (5 μg/mL).

compound	conc. (μM)	mean ± SEM of claudin-1 expression (mean fluorescence intensity)a	mean ± SEM of occluding expression (mean fluorescence intensity)a
1	50	17.12 ± 0.74	31.34 ± 1.27***
	25	14.18 ± 1.04	19.18 ± 0.65*
	12.5	6.86 ± 0.43	13.27 ± 0.24
2	50	34.13 ± 0.56***	32.25 ± 0.05***
	25	20.73 ± 0.36**	21.77 ± 1.69**
	12.5	14.41 ± 0.11	12.54 ± 0.18
4	50	22.83 ± 0.85***	25.71 ± 0.33***
	25	13.14 ± 1.41	13.83 ± 0.72
	12.5	5.01 ± 1.56	10.27 ± 1.11
5	50	27.21 ± 0.58***	34.73 ± 4.69***
	25	20.46 ± 0.24***	22.64 ± 2.58***
	12.5	16.74 ± 1.40	18.46 ± 0.00
6	50	34.97 ± 3.84***	31.92 ± 1.86***
	25	23.72 ± 1.49***	21.1 ± 0.49**
	12.5	12.24 ± 0.05	18.52 ± 0.30
8	50	36.50 ± 0.35***	32.27 ± 1.21***
	25	30.04 ± 0.06***	24.34 ± 0.44 ***
	12.5	19.29 ± 2.10*	17.80 ± 0.61
9	50	26.08 ± 3.04***	22.88 ± 2.84***
	25	16.35 ± 1.36	16.06 ± 1.32
	12.5	14.96 ± 2.23	10.89 ± 1.47
12	50	32.36 ± 6.64***	25.11 ± 0.67***
	25	24.86 ± 3.64***	18.73 ± 0.43
	12.5	8.09 ± 0.79	9.09 ± 2.63
13	50	33.58 ± 3.74***	24.77 ± 0.22***
	25	22.52 ± 1.71***	18.59 ± 0.23
	12.5	13.55 ± 1.34	8.82 ± 0.60
14	50	31.45 ± 0.64***	35.80 ± 0.37***
	25	22.27 ± 0.41***	23.48 ± 2.37***
	12.5	14.39 ± 1.35	18.89 ± 0.53
control		31.19 ± 0.77	30.17 ± 5.98
LPS		5.20 ± 2.95	7.94 ± 1.48

^aData are expressed as mean ± SEM of n = 3. ***, **, and * denote respectively p < 0.001, p < 0.01, and p < 0.05 vs LPS alone treated cells.

Further examination of compounds 1, 2, 4–6, 8, 9, and 12–14 for their inhibitory effect on IL-6 and TNF-α release showed that selected diterpenoids significantly inhibited the IL-6 release (Table 2, p < 0.001 vs LPS alone). However, compounds 5, 6, and 9 were the most active ones at all concentrations applied (Table 2). On the other hand, the majority of the compounds were able to inhibit TNF-α release (p < 0.001 vs LPS alone) in a concentration-dependent manner, and, among them, compound 5 was the most active (p < 0.001 vs LPS alone). During inflammation, oxidative stress also is of primary importance, and various intermediates are able to affect cell function.²⁶ Among them, nitrotyrosine is a marker of oxidative stress, and its levels were found to be elevated in refractory wounds.²⁷ Thus, the compounds were investigated for their effect on nitrotyrosine formation, as triggered by LPS in the HaCaT cells.³⁹ The results exhibited a notable inhibition of the nitrotyrosine formation by all compounds investigated in a concentration-dependent manner (p < 0.001 vs LPS alone). Other important regulators in wound-healing processes are tight-junction proteins. These proteins comprise three subgroups of transmembrane proteins, occludin, claudins, and junctional adhesion molecules that are present in epithelial and endothelial cells. These are responsible for the paracellular passage of molecules between adjacent cells.²⁸ Recent studies showed that the two tight-junction proteins occludin and claudin-1, which are present in the migrating epithelial cells at the wound edge, are absent in chronic wounds, supporting the pivotal role of tight junctions in wound-healing processes.²⁹ Therefore, the release inhibition of pro-inflammatory cytokines and an upregulation of tight-junction proteins could be considered as a promising strategy for the development of agents enhancing the wound-healing process. As shown in Table 3, under the experimental conditions utilized, the test compounds 1, 2, 4–6, 8, 9, and 12–14 significantly induced claudin-1 and occludin expression in a concentration-related manner in LPS-stimulated HaCaT cells (p < 0.001 vs LPS alone, Table 2), highlighting the possible ability of such diterpenoids in enhancing tissue regeneration in the wound-healing process, by acting on key mediators of this process such as cytokines, oxidative stress, and tight-junctions proteins.

Table 2. ¹³C NMR Spectroscopic Data of Compounds 1–5 in CHCl₃-d or MeOH-d₄ (150.92 MHz).

position	1	2	3	4	5	8	12	13a	14
1C	28.5	28.5	28.7	28.6	28.7	25.0	24.8	26.8	27.0
2C	25.5	25.3	25.3	25.2	25.5	25.2	25.1	26.3	25.3
3C	70.1	70.3	75.1	75.4	75.0	75.6	75.8	76.2	77.1
4C	41.2	41.2	37.6	37.5	37.6	38.1	38.1	38.9	39.2
5C	43.6	43.4	43.6	43.2	43.5	40.0	39.0	41.8	42.8
6C	34.9	34.9	43.8	34.9	34.9	38.8	38.7	39.5	76.2
7C	199.4	199.4	200.1	200.2	199.9	211.8	210.8	214.6	209.7
8C	130.7	131.3	131.2	131.3	131.3	51.2	49.7	51.3	45.0
9C	166.3	165.4	165.3	166.2	171.0	81.8	98.6	97.9	96.2
10C	40.6	40.4	40.8	40.5	40.8	43.2	42.7	44.0	42.9
11C	30.4	27.1	27.0	24.3	27.3	34.9	28.9	30.2	29.3
12C	24.4	24.5	26.5	27.0	24.9	21.7	37.8	40.4	38.6
13C	124.5	137.7	168.3	137.6	138.0	125.0	84.1	91.3	89.9
14C	110.7	144.1	117.9	144.0	142.3	110.8	44.0	47.2	46.2
15C	143.2	97.8	170.8	97.0	102.4	143.2	102.0	106.4	105.2
16C	138.8	171.3b	98.9	171.3	165.3	138.7	177.9	76.2	75.8
17C	11.6	11.8	11.6	11.9	11.7	8.4	9.3	9.3	9.2
18C	21.9	21.8	27.5	27.4	27.5	28.2	28.0	28.7	24.9
19C	67.2	67.1	21.8	21.7	21.8	21.9	21.6	22.1	27.9
20C	18.1	18.7	18.2	18.8	18.2	16.3	18.5	17.5	20.1
21C	171.2	171.2b			57.3		51.7	55.3	55.2
22C	21.0	21.1

^aMeOH-d₄ used as NMR solvent.

^bInterchangeable signals.

Labdane-type diterpenoids have gained attention over the past few decades due to their wide range of biological activities, such as antifungal, cytotoxic, and antimutagenic effects.^29,30 Previous studies have reported the anti-inflammatory activity of labdane diterpenoids through inhibition of the NF-κB pathway, consequently leading to a downregulation of the production of pro-inflammatory cytokines, e.g., IL-6, TNF-α, and the expression of iNOS and COX-2 at the mRNA and protein levels (pre- and post-translational) in LPS-stimulated macrophages.³¹ It has also been reported that labdane diterpenes bearing an α,β-unsaturated carbonyl group at the decalin unit or a γ-lactone ring (instead of a furan ring) possess high NF-κB inhibitory potential and, therefore, anti-inflammatory activity due to the covalent adduct formation with its Cys62 residue of the p50 subunit.^18,31 In one recent study, Khan et al. reported that compound 7 can suppress LPS-induced NF-κB activation, which resulted in the downregulation of iNOS and COX-2 protein expression, and in addition, it suppressed TNF-α production.¹⁸

Labdane-type diterpenoids have been described previously from different plant species. The most similar diterpenoids to those of R. persica were isolated mainly from Leonorus, Ballota, and Roylea.^17,19,32,33 However, the majority of those compounds from the above-mentioned sources were shown to possess no hydroxy groups at position C-3. In contrast, this study has demonstrated that all labdane-type diterpenoids obtained from R. persica possess a hydroxy group at C-3, pointing to the presence of a further oxidation step in the biosynthetic pathway of R. persica secondary metabolites.

In conclusion, the present phytochemical investigation of the aerial parts of R. persica has led to the isolation of 14 labdane-type diterpenoids, of which compounds 1–5, 8, and 12–14 were identified as new natural products. In order to determine the relative configuration of their ambiguous chiral centers, DP4+ NMR chemical shift probability calculations were applied, and the absolute configurations of the compounds were established using electronic circular dichroism and quantum chemical calculation methods. Furthermore, extracts and subfractions obtained from liquid–liquid extraction, along with selected isolated compounds, were examined for their activity on HaCaT cells in inflammatory conditions, by investigating their activity on IL-6 and TNF-α release and nitrotyrosine formation, along with their effect on claudin-1 and occludin expression. It may be concluded that the aerial parts of R. persica are a rich source of labdane-type diterpenoids with anti-inflammatory and potential wound-healing activity in keratinocytes, thus supporting the traditional application of this plant for the treatment of inflammation and wounds.^14,15

Experimental Section

General Experimental Procedures

Optical rotations were measured using a JASCO P-2000 polarimeter (Easton, MD, USA). UV and ECD spectra were measured on a JASCO J-1500 CD spectrometer. IR spectra were acquired on a Bruker Alpha FT-IR apparatus equipped with a Platinum ATR module (Bruker Daltonics, Bremen, Germany). One- and two-dimensional NMR experiments were recorded on a Bruker Avance II+ 600 spectrometer operating at 600.19 MHz (¹H) and 150.91 MHz (¹³C) at 300 K (chemical shifts δ in ppm, coupling constants J in Hz). Chloroform-d or methanol-d₄ was used as NMR solvent. HRESIMS analysis was performed on a Bruker micrOTOF-QII mass spectrometer after LC separation, using the following parameters: ESIMS parameters: 1:5 split from HPLC, dry temperature: 220 °C; dry gas (N₂): 6.00 L/min; nebulizer (N₂) 23.2 psi; full-scan mode: m/z 100–1500; ion polarity: positive or negative; capillary voltage: 3.5 kV; end plate offset: −0.5 kV. HPLC analysis was carried out on a Shimadzu LC-20AD XR system (Düsseldorf, Germany) equipped with a DAD detector, autosampler, and column thermostat.

LC parameters: stationary phase: Phenomenex Aqua C₁₈ 5 μm, 250 × 4.6 mm and a corresponding guard column; mobile phase: A = H₂O (+0.1% acetic acid in the case of LC-MS), B = acetonitrile; gradient: 0 min: 2% B; 30 min: 98% B; 40 min: 98% B; 40.1: stop; temperature: 35 °C, flow: 0.8 mL/min, injection volume: 5–15 μL, sample concentration = 2 mg/mL in CH₃CN or THF.

Plant Material

The aerial parts of Rydingia persica (Burm.f.) Scheen & V.A.Albert were collected in Hajiabad, Hormozgan Province, Iran, in March 2016 and identified by MSc. Rahman Adaspour (botanist). The plant material was shade-dried and stored until being used. A specimen (voucher number 1042) is deposited at the herbarium of the Agricultural and Natural Resources Research and Training Center of Hormozgan Province, Hormozgan, Iran.

Extraction and Isolation

A 500 g aliquot of the aerial parts of R. persica was milled and extracted with MeOH (3 L) in an ultrasonic bath for 10 min. The procedure was repeated eight times, and filtrates of each extraction were combined. Solvent evaporation was conducted under a vacuum using a rotatory evaporator (40 °C), resulting in 73 g of a crude methanolic extract. A part of the dried extract (70 g) was suspended in 200 mL of water and extracted with petroleum ether, diethyl ether, ethyl acetate, and n-butanol (9 × 200 mL each) to afford after evaporation 25.0, 0.3, 4.9, and 16.5 g of each subfraction, respectively, as well as 16.4 g of an aqueous subfraction. A 17.5 g quantity of the petroleum ether subfraction was subjected to chromatography on a silica gel column (Ø = 7 cm, l = 35 cm), using gradient elution with petroleum ether–EtOAc, from 0% to 100% EtOAc in 10 steps (600 mL, 10% increase in each step). Elution was continued using 5%, 10%, and 50% of MeOH in EtOAc (600 mL each). The fractions obtained were analyzed by TLC and subsequently combined to afford 28 pooled fractions. Furthermore, all fractions were analyzed using HPLC-DAD. F11 (1.1 g) was loaded on a Sephadex LH-20 column and isocratically eluted with CH₂Cl₂–acetone (85:15, v/v), resulting in 49 subfractions (F11-S1 to F11-S49). Among them, F11-S7 was identified as a pure compound (7, 44 mg). F11-S8 (290 mg) was loaded on a silica gel column (Ø = 1.5 cm, l = 40 cm). Subsequent isocratic elution with CH₂Cl₂–acetone (75:25, v/v) resulted in the isolation of F11-S8-P12 (4, 7 mg). Subfraction F11-S12 (70 mg) was purified using a silica gel column (Ø = 1 cm, l = 40 cm), eluted isocratically with CH₂Cl₂–acetone (95:5, v/v), resulting in the isolation of F11-S12-P10 (9, 2.1 mg) and F11-S12-P13 (11, 3.6 mg). F11-S13 (43 mg) was subjected to silica gel column (Ø = 0.5 cm, l = 25 cm) chromatography and eluted isocratically with CH₂Cl₂–petroleum ether–acetone (5:4:1, v/v), resulting in the purification of F11-S13-P15 (6, 1.4 mg). Subsequently, F11-S23 (48 mg) was purified by silica gel column (Ø = 0.5 cm, l = 25 cm) chromatography using isocratic elution with CH₂Cl₂–acetone (74:26, v/v), to afford F11-S23-P28 (3, 12.6 mg). F13 (1.4 g) was subjected to Sephadex LH-20 column chromatography (Ø = 2 cm, l = 100 cm), eluted with CH₂Cl₂–acetone (85:15, v/v), to afford 60 subfractions (S1 to S60). S9 (84 mg) was applied on a silica gel column (Ø = 1 cm, l = 40 cm) and eluted isocratically with CH₂Cl₂–acetone (85:15, v/v), which led to the isolation of F13-S9-P5 (1, 19 mg). F13-S12 (45 mg) was subjected to chromatography over a silica gel column (Ø = 1 cm, l = 25 cm), which was eluted with CH₂Cl₂–acetone (82:18, v/v), resulting in the isolation of F13-S12-P10 (8, 5 mg). F13-S14 (25 mg) was purified using a silica gel column (Ø = 0.5 cm, l = 20 cm), eluted with CH₂Cl₂–acetone (80:20, v/v), to afford F13-S14-P13-15 (12, 5 mg). F13-S16 (20 mg) was applied on a silica gel column (Ø = 0.5 cm, l = 25 cm) and eluted with CH₂Cl₂–MeOH (95:5, v/v), which led to the isolation of F13-S16-P14-17 (2, 5 mg). Additionally, F13-S18 (25 mg) was loaded on a silica gel column (Ø = 0.5 cm, l = 25 cm) and eluted with CH₂Cl₂–acetone (85:15 v/v) to afford F13-S18-P16-20 (10, 5 mg). F16 (500 mg) was subjected to Sephadex LH-20 column chromatography (Ø = 2 cm, l = 100 cm) and eluted with CH₂Cl₂–acetone (85:15 v/v), which resulted in 43 subfractions (F16-S1 to F16-S43). Subfraction F16-S5 (28 mg) was purified using a silica gel column (Ø = 0.5 cm, l = 25 cm), eluted with CH₂Cl₂–acetone (85:15, v/v), resulting in the isolation of F16-S5-P21-25 (13, 1.6 mg). Furthermore, F16-S7 (39 mg) was subjected to silica gel column chromatography (Ø = 1 cm, l = 25 cm) and eluted with CH₂Cl₂–acetone (90:10, v/v) to afford F16-S7-P23-26 (5, 4.2 mg). Additionally, F16-S13 (14 mg) was purified using a silica gel column (Ø = 0.5 cm, l = 20 cm), eluted with CH₂Cl₂–acetone (80:20, v/v), to afford F16-S13-P27-40 (14, 3 mg).

Persicunin A (1):

solid, amorphous gum; [α]_D²⁰ +28.3 (c 0.1, CHCl₃); UV (CH₃CN) λ_max (log ε) 208 (2.37), 247 (2.54) nm; ECD (CH₃CN) λ_max (Δε) 209 (+17.93), 245 (−4.91), 305 (−1.51), 347 (+1.88); IR ν_max 3955, 2925, 2856, 1730, 1666, 1455, 1171 cm^–1; ¹H and ¹³C NMR (CDCl₃), see Tables 1 and 2; HRESIMS m/z 375.216 (calcd for C₂₂H₃₁O₅, 375.2166, Δ 0.8 ppm).

Persicunin B (2):

solid, amorphous gum; UV (CH₃CN) λ_max (log ε) 200 (2.95), 246 (2.75) nm; ECD (CH₃CN) λ_max (Δε) 217 (+19.77), 243 (−6.60), 302 (−1.02); IR ν_max 2955, 2925, 2856, 1731, 1460, 1171 cm^–1; ¹H and ¹³C NMR (CDCl₃), see Tables 1 and 2; HRESIMS found m/z 407.2059 (calcd for C₂₂H₃₁O₇, 407.2064, Δ 1.4 ppm).

Persicunin C (3):

pale yellow, amorphous gum; UV (CH₃CN) λ_max (log ε) 247 (2.13) nm; ECD (CH₃CN) λ_max (Δε) 212 (+6.42), 246 (−3.14); IR ν_max 3372, 3002, 3959, 2875, 1753, 1643, 749 cm^–1; ¹H and ¹³C NMR (CDCl₃), see Tables 1 and 2; HRESIMS found m/z 349.2008 (calcd for C₂₀H₂₉O₅, 349.2010, Δ 0.5 ppm).

Persicunin D (4):

off-whitish, amorphous powder; UV (CH₃CN) λ_max (log ε) 200 (3.51), 247 (3.47) nm; ECD (CH₃CN) λ_max (Δε) 216 (+30.77), 246 (−10.70), 302 (−1.32); IR ν_max 2956, 2927, 2857, 1732, 1460, 1172 cm^–1; ¹H and ¹³C NMR (CDCl₃), see Tables 1 and 2; HRESIMS found m/z 347.1872, calcd for C₂₀H₂₇O₅, 347.1864, Δ −2.4 ppm).

Persicunin E (5):

colorless gum; UV (CH₃CN) λ_max (log ε) 247 (3.30) nm; ECD (CH₃CN) λ_max (Δε) 214 (+8.71), 245 (−4.76); IR ν_max 3481, 3086, 2938, 2873, 1764, 1651, 1023, 751 cm^–1; ¹H and ¹³C NMR (CDCl₃), see Tables 1 and 2; HRESIMS found m/z 363.2176 (calcd for C₂₁H₃₁O₅, 363.2166, Δ −2.7 ppm).

Calyonol (8):

off-whitish, amorphous powder; [α]_D²⁰ −37.6 (c 0.01, CHCl₃); UV (MeOH) λ_max (log ε) 202 (3.06) nm; ECD (CH₃CN) λ_max (Δε) 293 (−5.83); IR ν_max 2955, 2971, 2871, 1707, 1175, 1067, 1023, 873 cm^–1; ¹H and ¹³C NMR (CDCl₃), see Tables 1 and 2; HRESIMS found m/z 357.2033 (calcd for C₂₀H₃₀O₄Na, 357.2036, Δ 1.0 ppm).

Persicunin F (12):

off-whitish, amorphous powder; [α]_D²⁰ +30.2 (c 0.1, CHCl₃); UV (CH₃CN) λ_max (log ε) 208 (3.66), 248 (3.31), 313 (2.86) nm; ECD (CH₃CN) λ_max (Δε) 203 (−12.08), 228 (+15.28), 296 (−9.58); IR ν_max 2931, 1775, 1708, 1123, 1075, 1054, 1018 cm^–1; ¹H and ¹³C NMR (CDCl₃), see Tables 1 and 2; HRESIMS found m/z 365.1936 (calcd for C₂₀H₂₉O₆, 365.1970, Δ 9.1 ppm).

Persicunin G (13):

off-whitish, amorphous powder; [α]_D²⁰ +47.6 (c 0.03, CHCl₃); UV (MeOH) λ_max (log ε) 200 (3.17) nm; ECD (MeOH) λ_max (Δε) 205 (−5.41), 290 (−2.11); IR ν_max 2952, 2918, 2851, 1705, 1561, 1218, 1038, 771 cm^–1; ¹H and ¹³C NMR (MeOH-d₄), see Tables 1 and 2; HRESIMS found m/z 389.2294 (calcd for C₂₁H₃₄O₅Na, 389.2298, Δ 1.1 ppm).

Persicunin H (14):

off-whitish, amorphous powder; [α]_D²⁰ −20.2 (c 0.1, CHCl₃); UV (CH₃CN) λ_max (log ε) 247 (2.43) nm; ECD (CH₃CN) λ_max (Δε) 213 (+5.31), 310 (−3.35); IR ν_max 3480, 2979, 2927, 2879, 1708, 1035, 754 cm^–1; ¹H and ¹³C NMR (CDCl₃), see Tables 1 and 2; HRESIMS found m/z 405.2228 (calcd for C₂₁H₃₄O₆Na, 405.2248, Δ 4.8 ppm).

¹H and ¹³C NMR and DP4+ and ECD Calculations

Compounds were drawn in Maestro (Schrödinger Ltd., release 2018) and subjected to conformational analysis in MacroModel 09, using OPLS-3 as force field in the gas phase. The conformers obtained in an energy window of 5 kcal·mol^–1 were further applied to geometrical optimization and minimization using the DFT/6-31G(d,p) level in the gas phase in Gaussian 09/16.³⁴ Frequencies were calculated at the same level, and no imaginary frequencies were observed. Subsequently, NMR chemical shift calculations were conducted using a gauge-independent atomic orbitals (GIAO) method at the rmpw1pw91/6-31+G(d,P)/CPCM level in chloroform (12 and 14) or in MeOH (13). The shift tensors obtained were adjusted further to chemical shifts by using TMS proton and carbon chemical shifts, which were calculated with the same method. All chemical shifts were Boltzmann-averaged, and unscaled chemical shifts were used for DP4+ probability calculation based on the method published by Grimblat et al.³⁵ In order to calculate ECD spectra, optimized conformers obtained from the first optimization step were subjected to ECD calculation at the TD-DFT/B3LYP/6-31G(d,p)/CPCM or TD-DFT/cam-B3LYP/6-31G(d,p)/CPCM level in acetonitrile or MeOH. ECD curves were simulated on the basis of rotatory strengths using SpecDis v1.7³⁶ with a half-band of 0.2–0.4 eV and a UV shift of ±25 nm. All obtained spectra were Boltzmann-averaged prior to comparison with experimentally obtained ECD spectra.

Cell Culture

HaCaT is a spontaneously transformed aneuploid immortal keratinocyte cell line from adult human skin. This nontumorigenic cell line was cultured using Dulbecco’s modified Eagle’s medium (4 g/L glucose) supplemented with 10% (v/v) fetal bovine serum (FBS) and 1% (v/v) penicillin–streptomycin.

Antiproliferative Activity

Antiproliferative activity was evaluated using a colorimetric assay with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), as formerly reported.³⁷ HaCaT cells (5 × 103 cells/well) were plated on 96-well plates and allowed to adhere for 24 h at 37 °C in a 5% CO₂ atmosphere. Thereafter, the medium was substituted with either fresh medium alone or one containing serial dilutions of pure compounds (100–10 μM) or extract and subfractions (100–10 μg/mL) and incubated for 24 h. MTT (5 mg/mL) was then added to HaCaT cells. After 3 h, cells were lysed with 100 μL of a solution containing 50% (v/v) N,N-dimethylformamide and 20% (w/v) sodium dodecyl sulfate (SDS; pH = 4.5). A microplate spectrophotometer reader (Titertek Multiskan MCC/340-DASIT, Cornaredo, Milan, Italy) was used to measure the optical density (OD) in each well. The antiproliferative activity was calculated as % viability: 100 – [(OD treated/OD control) × 100].

Cell Treatment

HaCaT cells were plated as described before and, after adhesion, treated with extract or subfractions (100–10 μg/mL) or compounds (50–12.5 μM) for 1 h alone and then in the presence of lipopolysaccharides from E. coli (LPS; 5 μg/mL) for different experimental times, according to the cellular mediator to be evaluated.

IL-6 and TNF-α Determination

The IL-6 and TNF-α levels were assessed with an enzyme-linked immunosorbent assay (ELISA), as previously reported.³⁸ HaCaT cells were plated into 96-well plates (5 × 103 cells/well) and allowed to adhere for 24 h. Cells were then treated as previously indicated, for 24 h. Supernatants from HaCaT cells were then collected, and a commercial kit (Diaclone, Besançon, France) was used to perform the ELISA, according to the manufacturer’s instructions. Results were expressed as pg/mL.

Nitrotyrosine Formation and Claudin 1 and Occludin Expression by Cytofluorimetry

HaCaT cells were plated into 96-well plates (5 × 10³ cells/well) and treated with compounds (50–12.5 μM) or extract or subfractions (50–12.5 μg/mL) for 24 h, as previously indicated. For this analysis, HaCaT cells were collected and washed with phosphate-buffered saline (PBS). Fixing solution was added to cells for 20 min, and then HaCaT cells were incubated in fix perm solution (ThermoFisher Scientific, Waltham, MA, USA) for a further 30 min. Anti-nitrotyrosine (Merck Millipore, Milan, Italy), anti-claudin 1 (ThermoFisher Scientific), or anti-occludin (ThermoFisher Scientific) antibodies were then added for 1 h. The secondary antibody, in fixing solution, was added to the HaCaT cells, and cell fluorescence was measured by a fluorescence-activated cell sorter (FACSscan; Becton Dickinson) and analyzed by Cell Quest software (version 4; Becton Dickinson, Milan, Italy) as previously reported.³⁹

Data Analysis

Data are reported as means ± standard error of the mean (SEM) values of at least three independent experiments, each in triplicate. Statistical analysis was performed by analysis of variance test, and multiple comparisons were made by Bonferroni’s test by using Prism 5 (GraphPad Software, San Diego, CA, USA). p-Values lower than 0.05 were considered as significant.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jnatprod.0c00360.

• Copies of spectroscopic data for new compounds 1–5, 8, and 12–14; ¹H and ¹³C NMR data of the known compounds; results of conformational analysis of the new isolates; results of DP4+ analysis for compounds 12–14; results of antiproliferative activity of the methanolic extract, subfractions, and isolated compounds of R. persica; results of inhibitory effects of methanolic extract, subfractions, and isolated compounds of R. persica on nitrotyrosine formation, IL-6 and TNF-α release, as well as occludin and claudin-1 expression (PDF)

The authors declare no competing financial interest.