Antiviral and anti-inflammatory activities of chemical constituents from twigs of Mosla chinensis Maxim

1 Introduction

Influenza viruses had high pathogenicity and infectiousness, and is an important risk factor for human health. It had been exhibited the ability to invade the epithelial cells of the respiratory tract for the happening of the inflammation, and thereby result in influenza with the symptom such as fever, headache, and muscle pain. Influenza is one of the most common respiratory diseases. If the patients had not effective medical interventions, it could induce serious complications such as pneumonia, acute lung injury and even pulmonary fibrosis [1, 2]. Influenza viruses induced diseases had been become a worldwide public health problem and the main treatment is vaccine or drug. However, because of the extraordinary high rate of virus mutation and the side effects of existing drugs, it's essential to find ingredients with high effect and low toxicity from natural food. Phenolic compounds containing multiple phenolic hydroxyl groups, which can bind with targeting proteins of disease and possess significant activities of antioxidant, antiviral, and anti-inflammatory.

Mosla chinensis Maxim, recorded in Chinese Pharmacopoeia, is a medicinal and edible plant, which mainly distributed in southern China [3]. It belongs to the Labiatae family, a tomentose and aromatic plant that traditionally has been used as an herbal drug to treat colds in wet summers and aversion to cold with fever [4]. The leaves of M. chinensis are widely used as vegetable, herbal tea, beverage or food additives because of its human beneficial properties in China. Furthermore, M. chinensis is a productive source of essential oil and flavonoids. Several investigations have shown that essential oil of M. chinensis have the activities of antioxidant and antimicrobial [5-9] and the flavonoids exhibited the activities of anti-influenza A virus [10]. In our continuous search for compound of anti-influenza virus [11, 12], we found that few studies on the antiviral activity of other compounds isolated from M. chinensis were carried out. So, this paper was focus on exploring the activity of phenolic compounds. In this study, we investigated the extraction, structural analysis, biological activities and their possible mechanism researches of 35 compounds (covering 7 undescribed compounds and 28 known compounds) in the M. chinensis twigs.

2 Results and discussion

2.1 Structure characterization of the isolated compounds from M. chinensis

A comprehensive phytochemical investigation resulted in the isolation and identification of 35 compounds including seven new compounds (1–3, 19, and 27–29) and twenty-eight known compounds. The known compounds were listed as follows: 8-(4ʹʹ-hydroxyphenyl)-5,7,4ʹ-trihydroxyflavone (4) [13], apigenin-7-O-glucuronide methyl ester (5) [14], acacetin-7-O-glucuronide methyl ester (6) [15], Isolinariin B (7) [16], acacetin 7-O-[6ʹʹʹ-O-acetyl-β-D-galactopyranosyl-(1 → 3)]-β-D-xylopyranoside (8) [17], luteolin (9) [18], apigenin 7-O-β-D-glucopyranoside (10) [19], apigenin 4ʹ-O-β-D-glucopyranoside (11) [19], acacetin 7-O-β-D-xylopyranoside (12) [20], 4ʹ,5,7-trihydroxy-3ʹ,5ʹ-dimethoxyflavone 7-O-[β-D-apiofuranosyl (1ʹʹʹ → 2ʹʹ)]-β-D-glucopyranoside (13) [21], acacetin 7-O-β-D-apiofuranosyl-(1ʹʹʹ → 2ʹʹ)-β-D-glucopyranoside (14) [22], Diosmetin 7-O-β-D-xylopyranoside (15) [23], acacetin 7-O-[4ʹʹʹ-O-acetyl-β-D-apiofuransyl-(1ʹʹʹ → 3ʹʹ)]-β-D-xylopyranoside (16) [24], Sakuranetin (17) [25], Pyrroside A (18) [26], methyl lithospermate (20) [27], dimethy lithospermate (21) [27], hyprhombins A (22) [28], 3-(3ʹʹʹ, 4ʹʹʹ-dihydroxyphenyl)-acrylic acid 1-(3ʹʹ, 4ʹʹ-dihydroxyphenyl)-2-methoxycarbonylethyl ester (23) [29], sebestenoids C (24) [30], methyl salvianol acid C (25) [31], agrimonolide 6-O-β-D-glucopyranoside (26) [32], 3ʹ-hydroxyphenyl-3,4,5-trimethylgallate (30) [33], 4-[[(4-hydroxybenzoyl)oxy]methyl]phenyl-β-D-glucopyranoside (31) [34], 4-O-β-D-glucopyranosylbenzyl-3ʹ-hydroxyl-4ʹ-methoxybenzoate (32) [34], 4-[[(2ʹ, 5ʹ-dihydroxybenzoyl)oxy]methyl]phenyl-O-β-D-glucopyranoside (33) [19], amburoside A (34) [35], 4-hydroxybenzyl alcohol 4-O-[5-O-(4-hydroxy)benzoyl]-β-D-apiofuranosyl (1 → 2)-β-D-glucopyranoside (35) [36].

Compound 1, yellow amorphous powder, gave the molecular formula of C₂₆H₂₈O₁₃ based on its HR-ESI–MS ([M + Na] ⁺ m/z 571.1408, calcd. 571.1428). The ¹H NMR spectroscopic data (Table 1) of 1 showed the signals for six aromatic protons at δ_H 8.09 (2H, d, J = 8.6 Hz, H-2ʹ/6ʹ), 7.14 (2H, d, J = 8.6 Hz, H-3ʹ/5ʹ), 6.83 (1H, s, H-8), 6.41 (1H, s, H-6), an olefinic proton at δ_H 6.97 (1H, s, H-3), a hydroxyl at δ_H 12.96 (1H, s), and one methyl group at δ_H 3.87 (3H, s, H-OMe). Two anomeric protons δ_H 5.35 (1H, d, J = 6.7 Hz, H-1ʹʹʹ), 5.18 (1H, d, J = 7.5 Hz, H-1ʹʹ) were observed, which imply the presence of two aglycons. Acid hydrolysis afforded two sugar components as detected by the coupling constant values (J_{H-1ʹʹ, H-2ʹʹ}) and the GC analysis as β-D-xylose and β-D-apiose (Additional file 1: Fig. S60). The ¹³C NMR spectrum of 1 (Table 1) showed characteristic signals for the flavonoid skeleton at δ_C 182.5 (C-4), 164.2 (C-2), 99.8 (C-6), and 95.0 (C-8). The pyranose from of the sugars was also revealed from the ¹³C-NMR chemical shift values (Table 1) [37-39]. Based on ¹H, ¹³C NMR and HSQC data, the signals at δ_C 163.2 (C-4ʹ), 162.9 (C-7), 161.3 (C-5), δ_H 6.83 (1H, s, H-8), and 6.41 (1H, s, H-6) shows that 1 is an 5,7,4ʹ-trisubstituted flavonoid.

The ¹H and ¹³C NMR spectroscopic data of 1 (Table 1) is highly similar to acacetin [40], except for the additional presence of sugar moiety in 1. The above observation indicated 1 was glycoside derivative of acacetin. The position of glycosyl junction was identified by HMBC map. A series of HMBC correlations from H_Xyl-1 (δ_H 5.18) to C-7 (δ_C 162.9), from H_Api-1 (δ_H 5.35) to C_Xyl-4 (δ_C 69.9), from H_Xyl-4 (δ_H 3.43) to C_Api-1 (δ_C 109.3), enable the sugar chain of C-7 to be assigned as 7-O-[β-D-apiofuransyl-(1ʹʹʹ → 4ʹʹ)]-β-D-xylopyranoside. Thus, the structure of 1 was elucidated as acacetin 7-O-[β-D-apiofuransyl-(1ʹʹʹ → 4ʹʹ)]-β-D-xylopyranoside.

Compound 2 was purified as a yellow amorphous powder with the molecular formula of C₃₁H₃₄O₁₆ according to the HR-ESI–MS spectrum ([M + H] ⁺ m/z 663.1910, calcd. 663.1920). The ¹H and ¹³C NMR spectroscopic data of 2 (Table 1) were highly analogue to those of 1 except the presence of two acetyl groups at δ_C 170.5, 170.7, 20.7, and 21.1 in 2, as well as the minor change of chemical shifts in two sugars. Hence, 2 was deduced to be the acylated derivative of 1. Acid hydrolysis demonstrated the glycosidic nature of 2, which was identified as the β-D-glucose and β-D-apiose by the GC analysis and coupling constant values (J_{H-1ʹʹ, H-2ʹʹ}) (Additional file 1: Fig. S61). The position of glycosyl junction was identified by the HMBC correlations from H_Glc-1 (δ_H 5.25) to C-7 (δ_C 163.0), from H_Api-1 (δ_H 5.29) to C_Glc-2 (δ_C 76.9), from H_Glc-2 (δ_H 4.04) to C_Api-1 (δ_C 108.5). Furthermore, the sequence of the acetyl groups was deduced to be connected to C_Glc-6 and C_Api-5 due to the HMBC correlations from H_Glc-6 (δ_H 3.79/3.20) to C-7ʹʹ (δ_C 170.5), from H_Api-5 (δ_H 3.66/3.52) to C-6ʹʹʹ (δ_C 170.7). Therefore, compound 2 was identified as acacetin 7-O-[5ʹʹʹ-O-acetyl-β-D-apiofuransyl-(1ʹʹʹ → 2ʹʹ)]-6ʹʹ-O-acetyl-β-D-glucoside (Fig. 1).

Open image in new window

Compound 3 was obtained as a yellow amorphous powder. It showed a quasi-molecular ion peak at m/z 579.1717 [M + H] ⁺ (calcd. 579.1714) in the HR-ESI–MS data, suggesting a molecular formula C₂₇H₃₀O₁₄. The ¹H and ¹³C NMR spectroscopic data of 3 (Table 1) were also very similar to those of 1, apart from the extra methoxy group at C-3ʹ in 3 rather than hydrogen group at C-3ʹ in 1, which was supported by HMBC correlation between H-3ʹ-OMe (δ_H 3.67, s) and C-3ʹ (147.6) (Fig. 2). The anomeric configuration of D-xylose and D-apiose was confirmed to be β-configuration, according to the J value (J = 7.5 and 6.7 Hz) of the anomeric proton in the two sugar units (Additional file 1: Fig. S62). Thus, the structure of 3 was elucidated to be 3ʹ, 4ʹ-dimethoxyluteolin-7-O-[β-D-apiofuransyl-(1ʹʹʹ → 4ʹʹ)]-β-D-xylopyranoside.

Open image in new window

Compound 19, white amorphous powder, gave molecular formula of C₂₉H₂₆O₁₂ deduce from the HRESIMS spectrum (m/z 589.1293 [M + Na] ⁺, calcd. 589.1322). The ¹H NMR spectroscopic data (Table 2) of 19 showed signals for two olefinic methine protons at δ_H 7.58 (1H, d, J = 16.0 Hz, H-7), 6.37 (1H, d, J = 16.0 Hz, H-8), three oxymethine protons at δ_H 5.19 (1H, dd, J = 8.0, 5.0 Hz, H-8ʹ), 5.14 (1H, d, J = 5.0 Hz, H-7ʹʹ), and 5.07 (1H, d, J = 5.0 Hz, H-8ʹʹ), two methoxy protons at δ_H 3.69 (3H, s, H-9ʹʹ-OMe) and 3.62 (3H, s, H-9ʹ-OMe), two methylene protons at δ_H 3.03 (1H, dd, J = 13.5, 8.0 Hz)/3.31 (1H, dd, J = 13.5, 5.8 Hz), and nine phenyl protons. The spectra of ¹³C NMR and DEPT displayed twenty-nine carbon signals, covering twelve quaternary carbons [containing three carbonyls at δ_C 170.7 (C-9ʹʹ), 168.1 (C-9ʹ), 166.5 (C-9)], fourteen methines (including three oxymethines at δ_C 76.6, 75.8, and 73.4, two olefinic carbons at δ_C 145.9 and 116.1, and nine phenyl carbons), one methylene, and two methyls. A comprehensive analysis of the ¹H and ¹³C NMR spectral data (Table 2) revealed that 19 was a lignin, structurally comparable to clinopodic acid C [41] except for the lack of signals for two methoxy groups at C-9ʹ and C-9ʹʹ in clinopodic acid C (Fig. 2). The two methoxy groups were assumed to be connected to C-9ʹ and C-9ʹʹ, respectively, according to the HMBC correlations of H-9ʹʹ-OMe (δ_H 3.69) with C-9ʹʹ (δ_C 170.7) and H-9ʹ-OMe (δ_H 3.62) with C-9ʹ (δ_C 168.1). According to biogenetic considerations and key ROESY correlations (Fig. 3) observed between H-7ʹʹ/H-8ʹʹ, H-7ʹʹ/H-6ʹʹ and H-7ʹʹ/H-2ʹʹ suggested that H-7ʹʹ and H-8ʹʹ were located on the same side of the ring. The CD spectrum showed a negative Cotton effect at 244 nm (Fig. 4), suggesting that the absolute configuration of the benzodioxane moiety was established as 7ʹʹR, 8ʹʹR [41]. Therefore, compound 19 is identified as 9ʹ, 9ʹʹ-dimethyl clinopodic acid C.

Open image in new window

Open image in new window

Compound 27 was isolated as a colorless oil. The molecular formula of 27 was assigned as C₂₁H₂₈O₉ based on the HR-ESI–MS at m/z 447.1638 [M + Na]⁺ (calcd. 447.1631). Analyses of the 1D NMR data (Table 2) of 27 revealed signals for one ketone (δ_C 199.7), one trisubstituted benzene ring [δ_C 157.8, 133.7 and 133.5; δ_H 7.67 (d), 7.54 (d) and 7.45 (t)], and two methyls [δ_C 25.9 and 24.6; δ_H 1.10 (s) and 1.04 (s)]. In addition, one anomeric proton at δ_H 5.01 (1H, d, J = 7.7 Hz, H-1ʹ) was observed, which implies the presence of one aglycon. Acid hydrolysis of 27 afforded one sugar moiety, which was identified as the β-D-glucose by the GC analysis and coupling constant values (J_{H-1ʹ, H-2ʹ}) (Additional file 1: Fig. S63). The 1D NMR spectroscopic data of 27 (Table 2) were virtually identical to those of (3R, 4aR, 10bR)-3,10-dihydroxy-2,2-dimethyl-3,4,4a,10b-tetrahydro-2H-naphtho[1,2-b]-pyran-5H-6-one [42]. The major difference is the presence glycosyl group of located at C-10 in 27 as reinforced by the HMBC correlations from H_Glc-1 (δ_H 5.01) to C-10 (δ_C 157.8).

The relative configuration of 27 was determined by ROESY correlations (Fig. 3) of H-10b/H-4a, H-4a/H-4β, H-10b/H-4β, H-4α/H-3 observed, determined that the junction of B/C ring adopted a cis configuration, suggesting H-10b and H-4a were located on the same side of the ring C, and H-3 was the opposite. Hence, two stereoisomers were conceivable at that point, namely 3R, 4aS, 10bS and 3S, 4aR, 10bR. Subsequently, the absolute configuration of 27 was assigned as 3R, 4aS, and 10bS by comparison of the calculated and experimental ECD data in Fig. 5. Therefore, compound 27 was identified as a (3R, 4aS, 10bS)-2,2-dimethyl-3-hydroxy-10-O-β-D-glucoside-3,4,4a,10b-tetrahydro-2H-naphtho[1,2-b]-pyran-5H-6-one, named Mosla chinensis glycoside B1.

Open image in new window

Compound 28, white amorphous powder, possesses a molecular formula of C₂₀H₂₄O₇ concluded form the HR-ESI–MS spectrum ([M + H] ⁺ m/z 377.1592, calcd. 377.1595). The ¹H NMR spectrum (Table 2) of 28 revealed the signal for five aromatic protons at δ_H 7.25 (2H, s, H-2/6), 6.90 (1H, d, J = 8.2 Hz, H-3ʹ), 6.88 (1H, s, H-6ʹ), and 6.85 (1H, d, J = 8.2 Hz, H-4ʹ), two methylene groups at 4.48 (2H, t, J = 6.8 Hz, H-8ʹ) and 3.00 (2H, t, J = 6.8 Hz, H-7ʹ), and five methoxy protons at δ_H 3.85 (6H, s, H-3/5-OMe), 3.81 (3H, s, H-2ʹ-OMe), 3.79 (3H, s, H-4-OMe), and 3.78 (3H, s, H-5ʹ-OMe). Twenty carbon signals were totally observed in the spectra of ¹³C NMR and DEPT, covering eight quaternary carbons (including one ketone carbonyl at δ_C 167.5), five methyls, two methylenes (including one oxymethylene at δ_C 66.9), and five methines (Table 2). High similarity in the spectra of 1D NMR was found between 28 (Table 2) and 2-phenylethyl 2,4-dihydroxy-3-methylbenzoate [43]. The main difference between them is the presence of five methoxy groups at C-3, 4, 5, 2ʹ, 5ʹ in 28 rather than the two hydroxy groups at C-3, 5 and one methyl group at C-4 in 2-phenylethyl 2,4-dihydroxy-3-methylbenzoate. The five methoxy groups were assumed to be connected to C-3, 4, 5, 2ʹ, 5ʹ, respectively, according to the HMBC correlations of C-3-OMe (δ_H 3.85) with C-3 (δ_C 56.6), C-4-OMe (δ_H 3.79) with C-4 (δ_C 56.4), C-5-OMe (δ_H 3.85) with C-5 (δ_C 56.6), and C-2ʹ-OMe (δ_H 3.81) with C-2ʹ (δ_C 61.1). Therefore, the structure of 28 was assigned as 2ʹ,5ʹ-dimethoxyphenethyl 3,4,5-trimethoxybenzoate.

Compound 29, white amorphous powder, protonated a molecule peak at m/z 319.1537 [M + H] ⁺ (calcd 319.1540) corresponding to the molecular formula of C₁₈H₂₂O₅. The ¹H NMR spectrum of 29 exhibited six aromatic protons at δ_H 7.10 (2H, d, J = 8.1 Hz, H-6,5ʹ), 7.09 (2H, d, J = 8.1 Hz, H-5,6ʹ), 6.58 (1H, s, H-2ʹ), 6.33 (1H, s, H-2) and aliphatic protons at δ_H 5.50 (1H, s, H-7), 3.92 (1H, dd, J = 11.2, 3.7 Hz, H-8ʹa), 3.78 (1H, dd, J = 11.2, 2.9 Hz, H-8ʹb), 3.34 (2H, s, H-4ʹ-CH₂), 2.57 (1H, qt, J = 7.1, 3.2 Hz, H-7ʹ), 2.00 (3H, s) and 1.38 (3H, d, J = 7.1 Hz). The ¹³C NMR spectrum of 29 exhibited 18 carbon signals, which including six quaternary carbons [covering three oxygen-bearing sp³ carbons at 79.4 (C-7), 69.1 (C-8ʹ), 62.9 (C-4ʹ-CH₂)], eight methine, two methyl, and two methylene. Three ¹H-¹H COSY correlated systems of H-5/H-6; H-7ʹ/ H-8ʹ, and H-5ʹ /H-6ʹ were observed (Fig. 2). The results indicated that compound 29 was deduced to be a monoaromatic hydrocarbon and was structurally similar to that of 2-phenylethyl 2,4-dihydroxy-3-methylbenzoate [44]. The significant difference between them were substituent groups at C-2, 7, 3ʹ, 4ʹ, 7ʹ, there were hydrogen, hydroxy, hydroxy, hydroxymethyl and methyl at C-2, 7, 3ʹ, 4ʹ, 7ʹ of 29 rather than hydroxy, carbonyl, hydrogen, hydrogen and methyl. The deduction can be verified by the HMBC correlations of C-7ʹ (δ_H 2.75) with C-1ʹ (δ_C 138.8), C-7 (δ_H 5.50) with C-1 (δ_C 134.9), C-8ʹ (δ_H 3.92, 3.78) with C-7 (δ_C 79.4) and C-7ʹ (δ_C 31.6), C-7ʹ-Me (δ_H 1.38) with C-7ʹ (δ_C 36.1), C-4ʹ-CH₂ (δ_H 3.34) with C-4ʹ (δ_C 127.4). Therefore, compound 29 was identified as 4-(hydroxy(2-(3-hydroxy-4-(hydroxymethyl) phenyl) propoxy) methyl)-2-methylphenol.

2.2 Biological evaluation

2.2.1 Anti-influenza A virus activity

The activity of compounds (6–18, 20–26, 31–35) against the influenza virus was evaluated by using A/WSN/33/2009 (H1N1) infected MDCK cells. In comparison with the positive control oseltamivir with IC₅₀ = 6.85 µM, compound 20 exhibited significant inhibition effects of H1N1 (IC₅₀ = 20.47 µM); However, other compounds had no anti-influenza activity. The results of western blot analysis showed that 20 could dramatically reduce the nucleoprotein protein expression at 2, 5, and 8 h, indicating that 20 inhibits influenza virus infection by interfering with the beginning phase in the viral life cycle (Fig. 6). Furthermore, the nucleoprotein distribution in infected cells was observed by fluorescence microscopy (Fig. 7). It showed that after virus infection for 2 and 5 h, the virus population in the MDCK cells of the DMSO group was dramatically higher than that of the experimental group. This result further indicated that the influenza virus could be inhibited by compound 20.

Open image in new window

Open image in new window

Glycoprotein hemagglutinin (HA) of the influenza virus has been used as a potentially important target for developing anti-influenza drugs [45]. A hemagglutinin inhibition (HI) assay was designed to check if 20 could prevent virus attachment to the cells through disturbing the connection between HA and cellular receptors. The results showed that 20 might effectively promote erythrocyte agglutination at 40 μM (Fig. 8A and B). The results indicated that 20 could bind to influenza virus surface antigen HA1, inhibiting the early adsorption process of WSN. In addition, it was revealed that the docking sites with high-affinity underlying interaction could be intently related to residues ASN 68 and ARG 224 through Molecular docking (Fig. 8C). However, no binding sites between 20 and HA were observed on the receptor binding domain (RBD) of HA1 sialic acids. Therefore, it could be concluded that 20 exhibited the antiviral influence on A/WSN/33/2009 (H1N1) virus by targeting the hemagglutinin fusion machinery.

Open image in new window

2.2.2 Anti-inflammatory activity

The influenza virus can lead to an excessive immune response and induce the production of inflammatory cytokines, such as IL-1 and IL-6 [46]. So, it would be valuable if the drugs had both antiviral and anti-inflammatory activities. Hence, we used the cells (LPS-activated RAW 264.7) to evaluate the impact of compounds (6–18, 20–26, 31–35) on preventing NO production. It was found that compounds 9, 22, 23, and 25 rendered moderate activity in preventing NO production (IC₅₀ = 22.78, 20.47, 27.66, and 30.14 µM, respectively), in comparison with the positive control L-NMMA (IC₅₀ = 21.80 µM) (Table 3).

3 Experimental procedures

3.1 General experimental procedures

A Jasco digital polarimeter (DIP-370, purchased from JASCO Corporation, Tokyo, Japan) was employed to examine optical rotations. NMR spectra were monitored by a Bruker AV 600 MHz spectrometer using an internal standard (tetramethylsilane) (Bruker BioSpin Group, Germany). An API-QSTAR Pulsar (Applied Biosystem Corporation, Canada) was hired to achieve HR-ESI–MS and ESI–MS. A Shimadzu UV-2401 spectrometer (Beckman, Brea, USA) was implemented to obtain the UV spectra. Column chromatography with various gels was conducted, including 75 μM of ODS-C₁₈ (YMC Co., Ltd., Japan), 75–150 μM of MCI gel (GHP20P, Mitsubishi Chemical Corporation, Tokyo, Japan), 43–63 mm of LiChroprep RP-18 (Merck), 80–100 & 200–300 mesh of silica gels (Qingdao Marine Chemical Co., Ltd., China), and Sephadex LH-20 (Amersham Biosciences AB, Uppsala, Sweden). An Agilent 1260 liquid chromatography system (Agilent, USA) was implemented for semipreparative and analytical HPLC analysis on a semipreparative Zorbax SB-C₁₈ column (5 μm, 250 × 9.4 mm, 3 ml/min) and an analytical Zorbax SB-C₁₈ column (5 μm, 250 × 4.6 mm, 1 ml/min), respectively. TLC was run for monitoring collected fractions on silica gel GF₂₅₄ plates (Qingdao Marine Chemical Co., Ltd., China). The visualization of spots on the plates were conducted by using an ultraviolet lamp at the wavelength of 254 nm or by heating with H₂SO₄-EtOH (5%).

3.2 Plant material

The M. chinensis twigs were harvested in August 2019 from Honghe Hani and Yi Autonomous Prefecture (Yunnan, China), with the authentication of Dr. Jindong Zhong (Kunming University of Science and Technology). A voucher specimen (serial number: KMUST201903) was stored at the Department of Life Science and Technology.

3.3 Extraction and purification

The air-dried powdered of M. chinensis twigs (15 kg) was extracted with 70% acetone/H₂O by refluxing for 24 h (30 L × 3 times). After filtration and evaporation procedures, the extract (1286 g) was yielded and thoroughly dissolved in H₂O. The mixture was then extracted by petroleum ether, chloroform, ethyl acetate, and n-butanol, respectively. The ethyl acetate extract (145 g) was separated to four fractions (Fr. A–D) by silica gel column (20 × 300 cm) eluted with dichloromethane-methanol (1:0–0:1).

Fr. B (32.0 g) was subjected to four fractions (Fr. B-1–B-4) through MCI (90% MeOH/H₂O) and RP-18 eluting with MeOH/H₂O (30–100%). Fr. B-2 (1.5 g) was separated by silica gel column eluted with chloromethane-methanol (40:1–2:1) to give 9 (5.4 mg) and 23 (33.5 mg). Fr. B-3 (10.0 g) was separated to four fractions (Fr. B-3-1–B-3-5) by ODS C-18 (MeOH/H₂O, gradient 20%–100%). Fr. B-3-2 (1.3 g) was subjected to silica gel column eluted with dichloromethane-methanol (15:0–1:1) to obtain compounds 2 (3.2 mg) and 7 (4.2 mg). Compounds 5 (20.3 mg), 17 (2.7 mg, t_R = 21.2 min), 26 (4.4 mg, t_R = 12.5 min), and 34 (4.1 mg, t_R = 15.8 min) were obtained from Fr. B-3-3 (2.3 g) by Sephadex LH-20 (MeOH) and semi-preparative HPLC (85% MeOH-H₂O, 3 ml/min). Fr. B-3-4 (1.1 g) was purified by Sephadex LH-20 (CH₂Cl₂–MeOH, 1:1) and semi-preparative HPLC (73% MeOH–H₂O, 3 ml/min) to yield 24 (5.8 mg, t_R = 14.8 min), 25 (5.7 mg, t_R = 16.0 min), and 33 (5.3 mg, t_R = 10.6 min). Separation of Fr. B-3-5 (2.0 g) with silica gel column to obtain 22 (10.0 mg). Fr B-4 (1.0 g) was chromatographed over an ODS C-18 column (MeOH/H₂O, 30–100%) and semi-preparative HPLC (68% MeOH–H₂O, 3 ml/min) to give 1 (4.0 mg, t_R = 8.0 min), 8 (1.6 mg, t_R = 12.5 min), 11 (3.2 mg, t_R = 17.0 min), 12 (3.1 mg, t_R = 19.8 min), 31 (7.9 mg, t_R = 22.6 min), and 32 (10.8 mg, t_R = 16.0 min).

Fr. C (33.0 g) was fractioned by RP-18 (MeOH/H₂O, gradient 30–100%) to obtain subfractions Fr. C-1–C-4. Fr. C-3 (5.4 g) was applied ODS C-18 column chromatography eluted with MeOH/H₂O (30%–100%) and silica column (CH₂Cl₂-MeOH, 8:1–1:1) to give 10 (2.2 mg), 13 (9.2 mg), 20 (37.1 mg), and 35 (3.1 mg). Fr. C-4 (9.4 g) was chromatographed successively over silica gel (dichloromethane-methanol 10:1–1:1), Sephadex LH-20 gel (MeOH) and semi-preparative HPLC (65% MeOH–H₂O, 3 ml/min) to give compounds 19 (14.1 mg, t_R = 18.1 min), 21 (110.0 mg, t_R = 16.5 min), and 14 (3.4 mg, t_R = 13.2 min).

Fr. D (29.0 g) was fractioned via Sephadex LH-20 gel, eluting with MeOH to afford subfractions Fr. D-1–D-4. Fr. D-2 (1.2 g) was purified by silica gel column eluted with PE/EtOAc (15:1–1:1) and Sephadex LH-20 (MeOH), successively, yield compounds 3 (5.3 mg), 4 (6.2 mg), and 6 (7.3 mg). Compounds 15 (4.1 mg, t_R = 14.6 min), 16 (9.3 mg, t_R = 17.5 min), and 27 (7.2 mg, t_R = 23.8 min) were obtained from Fr. D-3 (4.2 g) by Sephadex LH-20 (MeOH) and semi-preparative HPLC (52% MeOH–H₂O, 3 ml/min). Fr. D-4 (11.2 g) was separated by silica gel column eluting with petroleum ether-ethyl acetate (10:1–1:1) to give Fr. D-4-1–D-4-3. Compound 28 (7.1 mg, t_R = 13.5 min) was obtained from Fr. D-4-2 (334.6 mg) by semi-preparative HPLC (85% MeOH–H₂O, 3 ml/min). Fr. D-4-3 (2.7 g) was applied to ODS column chromatography eluted with MeOH/H₂O (30–100%) and silica column eluting with chloromethane-methanol (30:1–1:1) to give 18 (12.1 mg), 29 (8.2 mg), and 30 (4.2 mg).

3.3.1 Acacetin 7-O-[β-D-apiofuransyl-(1ʹʹʹ → 4ʹʹ)]-β-D-xylopyranoside (1)

Amorphous powder; [α]25 D = − 25.1 (c = 0.10, MeOH); IR (KBr) v_max 3495, 3374, 2917, 2866, 1660, 1604, 1583, 1497, 1428, 1377, 1239, 1025, and 834 cm⁻¹; UV (MeOH): λ_max (log ε) 324 (2.23) nm; HRESIMS m/z 571.1408 [M + Na] ⁺ (calcd for C₂₆H₂₈O₁₃Na, 571.1428); ¹H and ¹³C NMR data (Table 1).

3.3.2 Acacetin 7-O-[4ʹʹʹ-O-acetyl-β-D-apiofuransyl-(1ʹʹʹ → 2ʹʹ)]-6ʹʹ-O-acetyl-β-D-glucoside (2)

Amorphous powder; [α]25 D = − 28.0 (c = 0.10, DMSO); IR (KBr) v_max 3430, 2921, 2850, 1728, 1615, 1587, 1489, 1365, 1300, 1182, 1080, 987, and 829 cm⁻¹; UV (MeOH): λ_max (log ε) 327 (3.23) nm; HRESIMS m/z 663.1910 [M + H] ⁺ (calcd for C₃₁H₃₄O₁₆, 663.1920); ¹H and ¹³C NMR data (Table 1).

3.3.3 3ʹ, 4ʹ-Dimethoxyluteolin 7-O-[β-D-apiofuransyl-(1ʹʹʹ → 4ʹʹ)]-β-D-xylopyranoside (3)

Amorphous powder; [α]25 D = − 35.0 (c = 0.10, DMSO); IR (KBr) v_max 3389, 2928, 2865, 1700, 1659, 1608, 1514, 1382, 1338, 1298, 1183, 1123, 1085, 987, and 830 cm⁻¹; UV (MeOH): λ_max (log ε) 326 (2.37) nm; HRESIMS m/z 579.1717 [M + H] ⁺ (calcd for C₂₇H₃₀O₁₄, 579.1714); ¹H and ¹³C NMR data (Table 1).

3.3.4 9ʹ,9ʹʹ-dimethyl clinopodic acid C (19)

White amorphous powder; [α]25 D = + 100.0 (c = 0.10, MeOH); IR (KBr) v_max 3431, 3040, 2954, 2925, 2851, 1741, 1607, 1585, 1506, 1440, 1269, 1117, 978, 858, and 812 cm⁻¹; UV (MeOH): λ_max (log ε) 327 (3.14) nm; HRESIMS m/z 589.1293 [M + Na] ⁺ (calcd for C₂₉H₂₆O₁₂Na, 589.1322); ¹H and ¹³C NMR data (Table 2).

3.3.5 Mosla chinensis glycoside B1 (27)

White amorphous powder; [α]25 D = -37.8 (c = 0.10, MeOH); IR v_max 3501, 3407, 3305, 2873, 1610, 1518, 1365, 1035, 826, and 714 cm⁻¹; ECD (MeOH) λ_max (Δε) 211 (-4.82), 247 (+ 2.43) nm; UV (MeOH): λ_max (log ε) 292 (2.57) nm; HRESIMS, m/z 447.1638 [M + Na] ⁺ (calcd for C₂₁H₂₈O₉Na_, 447.1631); ¹H and ¹³C NMR data (as shown in Table 2).

3.3.6 2,5-dimethoxyphenethyl 3,4,5-trimethoxybenzoate (28)

Colorless crystal; [α]25 D = − 12.3 (c = 0.10, MeOH); IR v_max 3445, 2933, 2843, 1715, 1592, 1512, 1460, 1415, 1334, 1228, 1127, 1028, 1002, 862, 807, and 764 cm⁻¹; UV (MeOH) λ_max (log ε) 310 (2.78) nm; HRESIMS, m/z 377.1592 [M + H] ⁺ (calcd for C₂₀H₂₄O₇, 377.1595); ¹H and ¹³C NMR data (as shown in Table 2).

3.3.7 4-(hydroxy(2-(3-hydroxy-4-(hydroxymethyl) phenyl) propoxy) methyl)-2-methy-lphenol (29)

White amorphous powder; [α]25 D = − 12.4 (c = 0.10, MeOH); IR v_max 3415, 2921, 2853, 1512, 1456, 1407, 1263, 1170, 1063, and 1023 cm⁻¹; UV (MeOH): λ_max (log ε) 251 (2.25) nm; HRESIMS, m/z 319.1537 [M + H] ⁺ (calcd for C₁₈H₂₂O₅, 319.1540); ¹H and ¹³C NMR data (as shown in Table 2).

3.4 Acid hydrolysis and determination of the absolute configuration of sugars

Based on the method reported by Wu et al. [47], the D glucopyranose configurations in compounds 1–3 and 27 were measured. Compounds 1–3 and 27 (1 mg per compound) were individually mixed with 3 ml HCl (2 M). Each mixture was boiled for 4 h at 100 ℃. After neutralization with NaHCO₃, the mixture was treated by EtOAc. Subsequently, the H₂O layer was evaporated and dissolved in DMSO (1.0 ml) before the acetic anhydride (40 μL) and 1-Methylimidazole (20 μL). After the reaction, extracted with EtOAc and analysed by GC. Monosaccharide compositions in compounds were identified by coeluting with authentic monosaccharide.

3.5 ECD quantification

The conformational structures of compounds were achieved from Chem3D modeling and ROESY spectra. In terms of the conformations, low-energy conformers of 27 were created via CONFLEX software by using an energy window (10 kcal/mol, MMFF94S) [48]. Density functional theory (DFT) method was employed to optimize the selected conformers in MeOH at the B3LYP/6-31 G (d) level [49]. Geometry optimizations and predictions of the conformersʹ ECD spectra were conducted by TD-DFT-B3LYP/6-311G (2d, p) level using a solvent (IEFPCM solvent model for methanol) [50]. SpecDis 1.71 was hired to generate the predicted curves of ECD, and Gaussian 16 package was applied to all predictions [51]. After UV correction, compound 27 spectrum was weighted using the Boltzmann distribution.

3.6 Anti-influenza virus assay

Based on the approach described by Dang et al. [52], anti-influenza virus assay was carried out. Briefly, prior to infection, MDCK cells (8 × 10³ cells/well) were cultivated for 24 h in 96-well plates, and the medium was removed. The mixture of compounds (at 3.125, 6.25, 12.5, 25, and 50 μM) and H1N1 virus was cultured at ambient temperature for 15 min and then transferred to the plates containing MDCK cells. The plates were stored at 37 ℃ with 5% CO₂ for 48 h. Subsequently, the antiviral activity was quantified using microscopy. The obtained antiviral activity was verified using the CellTiter-Glo luminescent cell viability assay (Promega, #G7570). Each compound's cytotoxicity was evaluated through incubating with uninfected MDCK cells for 48 h [53, 54].

3.7 Western blot assay

Influenza virus-infected MDCK cells were added with compound 20 (50 μM) at distinct time points (2, 5, 8, and 10 h). Cell lysates were harvested, and proteins from the supernatant were obtained [55]. After being electrophorized on a 12% SDS-PAGE electrophorized gel, protein extracts were distributed on a PVDF membrane and then cultivated for 1 h in blocking media (5% nonfat milk) at ambient temperature. Immunoblotting was performed using antibodies: anti-β-actin (SC-47778 (Catalog No.), Santa Cruz) and anti-nucleoprotein (EPR25683 (Abcam No.), GeneTex). The target proteins were visualized by chemiluminescence (ECL, Beyotime).

3.8 Immunofluorescence assay

MDCK cells (1 × 10⁵ cells) were loaded into each well of 24-well plates. The plates were kept under 5% CO₂ condition at 37 ℃. As the cells increased by 50%, the cells were added and cultured with A/WSN/33/2009 (H1N1) virus (MOI = 0.1) for 2 h. After the removal of the supernatant, the cells were rinsed two times using PBS. Subsequently, the 20 was supplemented to cells and stored under conditions of 5% CO₂ and 37 ℃. At 2, 5, 8, and 10 h of incubation, the cells were fixed using PFA in PBS (4%, Beyotime Biotechnology) at refrigerated temperature for 10 min. The fixed cells were then permeabilized for 10 min at room temperature using 0.1% Triton X-100 in PBS and blocked for 1 h at 37 ℃ using 3% BSA in PBS. Afterward, the treated cells were first stored at 4℃ overnight supplemented with nucleoprotein antibody diluted in 3% BSA (1:250, Abcam, CA, USA) and then cultured at ambient temperature for 1 h with fluorescein isothiocyanate (FITC)-labeled secondary antibody diluted in 3% BSA (1:250). The nucleus in cells was stained by DAPI for 10 min at ambient temperature. After staining, the fluorescence was examined by an inverted fluorescence microscope (Nikon A1R/A1, Shanghai, Japan) [56].

3.9 Hemagglutination inhibition assay

Hemagglutination inhibition (HAI) test was utilized to evaluate the activity of 20 against HA-mediated avian RBCs hemagglutination [57]. Briefly, 20 (10, 20, 40, and 80 μM) with influenza A/WSN/33/2009 (H1N1) (2^–6 hemagglutination titer) was supplemented into 96-well plates and cultured under an ambient condition for 1 h. Afterward, 1% chicken RBCs saline solution (50 μL) was loaded into each well. After incubation for 30 min at ambient temperature, the hemagglutination was examined.

3.10 Molecular docking

The protein structure of H1N1 HA (PDB ID: 6CFG) was achieved from the RCSB protein data bank. Chemdraw3D was hired to construct 20's 3D structures. The docking procedures were provided by the AutoDock Software with a graphics interface (AutoGrid/AutoDock 4.2.6 and Vina). With the AutoDock tools, the deletion of all water molecules was executed, while the refined model was added with the polar hydrogen atoms and charges. Then, docking was conducted using AutoDock/Vina based on the HA information and the grid box characteristics of the studied compound in the configuration file. In the process of docking, the 20 structure and HA protein structure were regarded as rigid [58, 59].

3.11 Nitric oxide production assay

Based on the previously reported approach [60, 61], NO production in cells was assayed. Cells were loaded into 96-well plates (8×10⁴/ well). The cells were treated by compounds (at 3.125, 6.25, 12.5, 25, and 50 μM) for 1 h, supplemented with LPS (1 μg/ml), and cultivated for 24 h. After treatment, a microplate reader (Thermo Fisher Scientific, Massachusetts, USA) was hired to quantify the absorbance values with the wavelength of 540 nm. The positive and negative controls were separately L-NMMA and DMSO.

4 Conclusions

This work systematically explored the phytochemical characteristics of thirty-five compounds extracted from M. chinensis twigs. The compounds included three unreported flavonoids (1–3), one undescribed phenylpropanoid (19) and three new monoaromatic hydrocarbons (27–29), and 28 known compounds. Compound 27's absolute configuration was interpreted and visualized with the assistance of ECD calculation. Compound 20 exhibited the most significant activity against A/WSN/33/2009 (H1N1) virus (IC₅₀ = 20.47 μM). Further research showed that 20 could bind to influenza virus surface antigen HA1 and inhibit the early adsorption process of the influenza A/WSN/33/2009 (H1N1) virus strain. Furthermore, compounds 9, 22, 23, and 25 displayed moderate inhibitory effects on the NO expression in LPS inducing Raw 264.7 cells with IC₅₀ values of 22.78, 20.47, 27.66, and 30.14 µM, respectively.

The effect of M. chinensis on influenza virus infection, controlling the adsorption of virus and excessive inflammatory reaction in the infection process. Our findings will enrich the study of the structural diversity of M. chinensis and provide insights into understanding the plant's anti-influenza function, which may launch scientific basis for the following research about the development of antiviral beverages and resources utilization of M. chinensis.

Notes

Acknowledgements

This paper was financed by National Natural Science Foundation of China (No. 31660100), Innovative Team of Yunnan Province (No. 2019HC018), the key Project of Yunnan Province (No. 202103AC10005, No. 202302AG050004), and Higher Educational Key Laboratory for New drugs for Viral Respiratory Diseases (Chinese Traditional Medicine) of Yunnan Province.

Author contributions

S.-Y. F. isolated and identified of the compounds; writing—original draft. N. J. was responsible for the biological activities assessment. J.-Y. Y. performed chemical calculation and wrote the paper. L.-Y. Y. and J.-C. D. contributed to the extraction, isolation, and identification of the compounds. D.L. and J.-D. Z. designed experiments. X.-Q. C. and R.-T. L. checked the whole manuscript. All authors read and approved the fnal manuscript.

Data availability

The datasets used or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare that there are no competing interests associated with this work.