Ursane Triterpenes and Norisoprenoids from Anchusa italica Retz. and Their Chemotaxonomic Significance

Abstract

A total of 31 compounds were isolated from the ethyl acetate and n-butanol fractions of Anchusa italica Retz., which contained one ursane triterpenoid, 2α,3β,19α-trihydroxy-23-formyl-urs-12-en-28,21β-olide (1), and five norisoprenoids: (2R,6R,9S)-9-hydroxy-4-megastigmen-3-one-2-O-β-D-glucopyranoside (3); (2R,6S,9S)-9-hydroxy-megastigman-4,7-dien-3-one-2-O-β-D-glucopyranoside (4); (+)-isololiolide β-D-glucopyranoside (5); (2S,8R)-loliolide β-D-glucopyranoside (6a); and (2R,8S)-loliolide β-D-glucopyranoside (6b). It also contained 25 known compounds (2 and 7–30). The chemical structures of the compounds, inclusive of their absolute configurations, were ascertained using spectroscopic methods such as NMR, HR-MS, and quantum chemical calculations (computational NMR and ECD), in combination with relevant literature data. Moreover, the chemotaxonomic significance of the isolated substances was discussed, with compounds 1, 2, and 7–13 potentially broadening the application of triterpenes as taxonomic markers for the classification of the genus Anchusa.

1. Introduction

Anchusa italica Retz. belongs to the genus Anchusa (Boraginaceae family) and is a perennial herbaceous plant that is naturally distributed in the Mediterranean region and tropical areas. In China, it is found in the northwestern and southwestern regions, such as Xinjiang, Gansu, and Sichuan [1]. As an important medicinal resource in the Uyghur medicine system, A. italica is recorded as “Gao Zi Wan” (meaning “tough”) and often used in the clinical treatment of cardiovascular and cerebrovascular diseases such as hypertension, palpitations, coughs, and depression [2]. Chemical investigations into A. italica have revealed that this plant contains abundant natural products, such as essential oils, triterpenoids, flavonoids, steroids, and alkaloids. From the perspective of pharmacological effects, the extracts and active compounds of A. italica exhibit a wide range of biological activities, including cardiovascular protection and anti-inflammatory, anti-bacterial, anti-oxidative, and neuroprotective effects [3]. To date, research has mainly focused on isolating and identifying total flavonoids and triterpenoid components, as well as exploring the mechanisms by which they protect myocardial cells [4]. Systematic analyses of their composition, activity, and contribution to plant taxonomy remain lacking.

To explore more compounds with biological activities and their contribution to the chemotaxonomic significance of this plant, we provide a certain basis for the development and utilization of A. italica. The whole plant powder was refluxed with 70% ethanol and then sequentially extracted with petroleum ether (PE), dichloromethane (CH₂Cl₂), ethyl acetate (EtOAc), and n-butanol (BuOH). The EtOAc and BuOH fractions were separated using chromatographic separation techniques: normal, reversed-phase silica gel column chromatography (CC), and HPLC. A total of thirty-one compounds (including nine triterpenes, twelve sesquiterpenes, four flavonoids, and six lignans) were identified, with six new compounds (1, 3, 4, 5, 6a, and 6b) as well as fourteen known compounds (15, 16, 18–25, and 27–30) first reported from A. italica (Figure 1). Compounds 5, 6a, and 6b are a differential isomer and a pair of stereoisomers, and in this study, the absolute configurations of these three compounds were identified for the first time using DP4+ analysis and ECD calculations. The chemotaxonomic significance of the isolated substances is also discussed.

2. Results

2.1. Structure Elucidation

Chemical investigation of the aerial parts of A. italica yielded six novel compounds (1, 3–5, 6a, and 6b) and 25 known analogues (2 and 7–30), including 2α,3β,19α-trihydroxy- 23-formyl-urs-12-en-28,21β-olide (1); (2R,6R,9S)-9-hydroxy-4-megastigmen-3-one-2-O- β-D-glucopyranoside (3); (2R,6S,9S)-9-hydroxy-megastigman-4,7-dien-3-one-2-O-β-D- glucopyranoside (4); (+)-isololiolide β-D-glucopyranoside (5); (2S,8R)-loliolide β-D- glucopyranoside (6a); (2R,8S)-loliolide β-D-glucopyranoside (6b); 2α,3β,21,24- tetrahydroxyoleanan-12-en-28-oic acid (2); niga-ichigoside F1 (7); niga-ichigoside F2 (8); pinfaensin (9); glucosyl tormentate (10); myrianthic acid (11); 24-epi-pinfaensic acid (12); hydroxyasiatic acid (13); (+)-vomifoliol (14) [3,4]; lippianoside E (15) [5]; 3-oxo-α- ionol-β-D-glucopyranoside (16) [6]; asysgangoside (17) [4]; (6S, 9R)-roseoside (18) [7]; sammangaoside B (19) [8]; (+)-isololiolide (20) [9]; tricin (21) [10]; 6-hydroxykaempferol 3-β-rutinoside (22) [11]; kaempferol 3-O-rutinoside (23) [12]; narcissin (24) [13]; (+)-syringaresinol (25) [14]; (+)-mediaresinol (26) [4]; dehydrodiconiferyl alcohol (27) [15]; vibruresinol (28) [16]; dehydrodiconiferyl alcohol 4-O-β-D-glucopyranoside (29) [17]; and syringaresinol-4′-O-β-D-glucopyranoside (30) [14].

Compound 1 was obtained as yellow amorphous powder with a molecular formula of C₃₀H₄₄O₆ deduced from m/z 501.3211 [M + H]⁺, with an unsaturation degree of 9. The 1D NMR data (

Table 1. ¹H and ¹³C-NMR data (500 and 125 MHz) of compounds 1 (CD₃OD, δ in ppm, J in Hz).

Position	δC	δH	Position	δC	δH
1	47.5	0.96, 1H, m 2.06, 1H, m	16	29.9	1.23, 1H, m 1.81, 1H, m
2	69.2	4.06, 1H, ddd, 5, 7, 10	17	45.8
3	82.6	3.11, 1H, d, 9.5	18	57.0	2.40, 1H, s
4	55.3		19	74.3
5	58.3	1.15, 1H, m	20	45.3	1.79, 1H, m
6	20.0	1.46, 1H, m 1.81, 1H, m	21	83.9	4.34, 1H, d, 5.5
7	34.0	1.61, 2H, m	22	34.0	1.61, 1H, m 3.34, 1H, d, 9.0
8	41.9		23	208.5	9.90, 1H, s
9	47.1	1.70, 1H, m	24	21.4	1.26, 3H, s
10	39.3		25	18.0	0.90, 3H, s
11	25.4	1.99, 1H, m 2.08, 1H, m	26	17.4	0.95, 3H, s
12	131.3	5.53, 1H, t, 4.0	27	26.2	1.39, 3H, s
13	137.7		28	185.4
14	42.7		29	13.8	1.05, 3H, d, 7.0
15	29.3	1.08, 1H, m 1.89, 1H, m	30	28.4	1.21, 3H, s

) revealed the presence of six tertiary methyl groups at [δ_H 0.90 (CH₃-25), 0.95 (CH₃-26), 1.05 (3H, d, J = 7 Hz, CH₃-29), 1.21 (CH₃-30), 1.26 (CH₃-24), and 1.39 (CH₃-27); δ_C 18.0, 17.4, 13.8, 28.4, 21.4, and 26.2, respectively]; three oxymethine protons at [δ_H 3.11 (1H, d, J = 9.5 Hz, H-3), 4.06 (1H, ddd, J = 5, 7, 10 Hz, H-2), and 4.34 (1H, d, J = 5.5 Hz, H-21); δ_C 82.6, 69.2, and 83.9]; one olefinic proton at [δ_H 5.53 (1H, t, J = 4 Hz, H-12); δ_C 131.3 and 137.7]; one aldehyde group at [δ_H 9.90 (1H, s, H-23); δ_C 208.5]; and one ester carbonyl group at (δ_C 185.4), suggesting that it is an ursane derivative. Its NMR data (Table 1) showed a strong similarity to 2α,3β,19α,23-tetrahydroxyurs-12-en-28, 21β-olide [3], except that the chemical shift of C-23 changed from δ_C 66.4 to δ_C 208.5 in compound 1, indicating that the C-23 changed from an oxymethylene to an aldehyde group.

The following correlations helped to confirm the presence of the ursolic acid skeleton: HMBC correlations of H-12 (δ_H 5.53) with C-11 and C-13 (δ_C 25.4 and 137.7); H-18 (δ_H 2.40) with C-13, C-17, and C-19 (δ_C 137.7, 45.8, and 74.3). Moreover, the following HMBC correlations helped to confirm the presence of the γ-lactone unit: correlations of H-21 (δ_H 4.35) with C-19, C-22, C-28, and C-30 (δ_C 74.3, 34.0, 185.4, and 28.4, respectively), together with the unsaturation degree of 9. Other correlations were also found, like H-23 (δ_H 9.90) with C-4 and C-24 (δ_C 55.3 and 21.4); CH₃-25 (δ_H 0.90) with C-1 and C-10 (δ_C 47.5 and 39.3); CH₃-27 (δ_H 1.39) with C-13 and C-14 (δ_C 137.7 and 42.7); and CH₃-30 (δ_H 2.40) with C-19, C-20, and C-21 (δ_C 74.3, 45.3, and 83.9). Based on biosynthesis considerations, the natural ursane skeleton has an α-orientation for H-5 and a β-orientation for H-18 and CH₃-25. The ROESY correlations of H-2 (δ_H 4.06) with H-18 (δ_H 2.40) and H-20 (δ_H 1.79), and H-5 (δ_H 1.15) with H-3 and H-21 (δ_H 3.11 and 4.34) suggested that H-2 is β-oriented, while H-3 and H-21 are α-oriented (Figure 2). Taking these data together, compound 1 was identified as 2α,3β,19α-trihydroxy-23-formyl-urs-12-en-28,21β-olide.

Compound 3 was obtained as a colorless oil, possessing a molecular formula of C₁₉H₃₂O₈ deduced from m/z 389.2170 [M + H]⁺ as well as the unsaturation degree of 4. The 1D NMR data (

Table 2. ¹H-NMR data (500 MHz) of compounds 3–6 (CD₃OD, δ in ppm, J in Hz) ¹.

Position	3	4	5	6a	6b
1			1.45, 1H, t, 12.0 2.65, 1H, dd, 2.0, 12.0	1.65, 1H, dd, 4.0, 14.0 2.65, 1H, dt, 2.5, 14.5	1.77, 1H, dd, 4.0, 14.0 2.67, 1H, dt, 2.5, 14.5
2	4.20, 1H, s	4.17, 1H, s	4.27, 1H, tt, 4.0, 12.0	4.28, 1H, m	4.30, 1H, m
3			1.38, 1H, t, 12.0 2.18, 1H, dd, 2.0, 12.0	1.55, 1H, dd, 4.0, 14.0 2.20, 1H, dt, 2.5, 14.5	1.44, 1H, dd, 4.0, 14.0 2.21, 1H, dt, 2.5, 14.5
4	5.81, 1H, d, 1.0	5.91, 1H, d, 1.0
5
6	2.10, 1H, m	2.77, 1H, d, 7.5	5.77, 1H, s	5.74, 1H, s	5.74, 1H, s
7	1.60, 1H, m 2.11, 1H, m	5.76, 1H, m
8	1.62, 2H, m	5.77, 1H, m
9	3.87, 1H, m	4.40, 1H, m	1.31, 3H, s	1.27, 3H, s	1.28, 3H, s
10	2.03, 3H, d, 1.0	1.92, 3H, d, 10.0	1.28, 3H, s	1.43, 3H, s	1.45, 3H, s
11	0.88, 3H, s	0.90, 3H, s	1.59, 3H, s	1.75, 3H, s	1.74, 3H, s
12	1.19, 3H, s	1.11, 3H, s
13	1.19, 3H, d, 6.0	1.30, 3H, d, 6.5
Glc
1	4.33, 1H, d, 8.0	4.36, 1H, d, 7.5	4.42, 1H, d, 8.0	4.37, 1H, d, 8.0	4.39, 1H, d, 8.0
2	3.13, 1H, m	3.17, 1H, m	3.12, 1H, m	3.17, 1H, m	3.17, 1H, m
3	3.33, 1H, m	3.33, 1H, m	3.29, 1H, m	3.36, 1H, m	3.35, 1H, m
4	3.24, 1H, m	3.29, 1H, m	3.24, 1H, m	3.27, 1H, m	3.27, 1H, m
5	3.24 1	3.22, 1H, m	3.29 1	3.27, 1H, m	3.27 1
6	3.65, 1H, m 3.83, 1H, m	3.67, 1H, m 3.83, 1H, m	3.63, 1H, m 3.89, 1H, m	3.64, 1H, m 3.85, 1H, m	3.65, 1H, m 3.86, 1H, m

¹ Overlapped with other signals.

and

Table 3. ¹³C-NMR data (125 MHz) of compounds 3–6 (CD₃OD, δ in ppm).

Position	3	4	5	6a	6b
1	43.2	42.8	45.6	42.9	45.4
2	77.2	77.4	72.9	74.2	74.5
3	201.2	200.9	48.9	46.9	44.4
4	124.0	125.0	35.9	37.1	37.0
5	168.6	164.7	183.8	185.8	185.7
6	54.5	58.6	113.4	113.1	113.1
7	25.9	128.3	173.8	174.4	174.4
8	38.2	138.5	88.3	88.9	88.9
9	75.4	77.0	30.1	30.9	31.0
10	24.6	23.7	25.1	26.6	26.7
11	21.0	25.5	25.4	27.1	27.0
12	24.5	21.0
13	19.8	21.0
Glc
1	102.0	102.5	102.7	102.9	103.2
2	75.1	75.3	74.8	75.3	75.3
3	78.1	78.1	77.8	78.5	78.4
4	71.7	71.5	71.5	71.7	71.7
5	77.8	78.0	77.8	77.9	77.9
6	62.9	62.6	62.6	62.8	62.8

) revealed the presence of four methyl groups at δ_H 2.03 (δ_C 24.6, CH₃-10), 1.19 (δ_C 21.0, CH₃-11), 1.19 (δ_C 19.8, CH₃-13) and 0.88 (δ_C 24.5, CH₃-12); one olefinic proton at [δ_H 5.81 (1H, d, J = 1.0 Hz, H-4) and δ_C 124.0]; a ketocarbonyl (δ_C 201.2) forming a conjugation with the olefinic bond; an isomeric proton at δ_H 4.20 (1H, s, H-2); and an oxygenated methyl proton at δ_H 3.87 (1H, m, H-9). Chemical shifts in the δ_H 3.23–3.90, along with an anomeric proton resonating at [δ_H 4.33 (1H, d, J = 8.0 Hz, H-1′), δ_C 102.0], were assigned as part of a sugar moiety and matched well with previously reported data as a β-glucopyranose. The ¹H and ¹³C NMR data (Table 2 and Table 3) for 3 were similar to those for (2R,6R,9R)-2,9-dihydroxy-4- megastigmen-3-one [18]. Only one more glucose unit is present in 3, and the chemical shift value of C-2 was down-field shifted from δ_C 76.0 (in (2R,6R,9R)-2,9-dihydroxy- 4-megastigmen-3-one) to δ_C 77.2 in 3. HMBC correlations were analyzed from CH₃-11 (δ_H 1.19) to C-1, C-2, C-6, and C-12 (δ_C 43.2, 77.2, 54.5, and 24.5, respectively); from H-10 (δ_H 2.03) to C-4, C-5, and C-6 (δ_C 124.0, 168.6, and 54.5); from H-7 (δ_H 2.11) and H-8 (δ_H 1.60) to C-6 (δ_C 54.5); and from H-13 (δ_H 1.19) to C-8 and C-9 (δ_C 38.2 and 75.4). Glucose was attributable to C-2 by HMBC correlations from δ_H 4.33 (H-1′) to δ_C 77.2 (C-2). The relative configuration of 3 was conducted to find correlations among H-2 (δ_H 4.20), H-6 (δ_H 2.10) and CH₃-11 (δ_H 0.88), providing β-oriented configurations.

To determine the absolute configuration of the sugar molecule of compound 3, acid hydrolysis followed by chemical derivatization was performed to determine the D/L configuration of β-glucopyranose (Figures S59 and S64). The β-glucopyranose from the acid hydrolysis of compound 3 and the standard β-D glucopyranose were chemically reacted to give the final product. Their retention times were compared by HPLC analysis, which showed that the β-glucopyranose in compound 3 was identified as the D-form, as they showed the same retention time (t_R = 20.2 min). To further verify the configuration of HO-9, two isomers (2R,6R,9S) and (2R,6R,9R) of 3 were subjected to NMR calculations with the DP4+ analysis. These results showed that (2R,6R,9S) of 3 (Figure S56) was the most likely structure for 3, with a 100.00% DP4+ probability based on all of the NMR data. The theoretical calculations of ECD spectra for 3 (2S,6S,9R) and the enantiomer (2R,6R,9S) were also performed using the TDDFT method. The calculated ECD spectra (2R,6R,9S) were in agreement with the experimental data (a positive peak at λ₂₅₀ nm and a negative peak at λ₃₀₀ nm), which indicated the absolute configuration (Figure 3). Taking these data together, compound 3 was identified as (2R,6R,9S)-9-hydroxy-4- megastigmen-3-one-2-O-β-D-glucopyranoside.

Compound 4 is a white powder, possessing a molecular formula of C₁₉H₃₀O₈ deduced from m/z 387.20135 [M + H]⁺. The unsaturation degree is 5. The 1D NMR data (Table 2 and Table 3) suggested compound 4 might be the glucoside derivative at C-2 of the known compound 2,9-dihydroxy-megastigman-4,7-dien-3-one, because the chemical shift value of C-2 was down-field shifted from δ_C 75.5 in 2,9-dihydroxy-megastigman- 4,7-dien-3-one to δ_C 77.4 in 4 [19]. Glucose was attributable to C-2 by HMBC correlations from H-1′ (δ_H 4.36) to C-2 (δ_C 77.4). The β-glucopyranose in 4 was identified as the D-form, as they showed the same retention time (t_R = 20.2 min) in HPLC analysis. The relative configuration was assigned by ROESY correlations of H-2 (δ_H 4.17) with H-8 (δ_H 5.77), H-9 (δ_H 4.40), and CH₃-11 (δ_H 0.90), suggesting that H-2, H-9, and CH₃-11 were on the same side of the molecule. The absolute configuration of 4 was determined via ECD calculations using the TDDFT method. The calculated ECD data for (2R,6S,9S) were in agreement with the experimental data of 4 (Figure 3), which indicated that compound 4 could be identified as (2R,6S,9S)-9-hydroxy-megastigman-4,7-dien-3-one-2-O-β-D- glucopyranoside.

Compound 5 was obtained as a colorless oil, possessing a molecular formula of C₁₇H₂₆O₈ deduced from m/z 381.1519 [M + Na]⁺ and ¹³C NMR data, as well as an unsaturation degree of 5. The 1D NMR data (Table 2 and Table 3) revealed the presence of three methyl groups at [(δ_H 1.59 (CH₃-11), 1.31 (CH₃-10) and 1.28 (CH₃-9); δ_C 25.4, 30.1, and 25.1)]; one olefinic proton at [δ_H 5.77 (1H, s, C-6), δ_C 113.4]; one ester carbonyl at (δ_C 173.8); and an isomeric proton at δ_H 4.27 (1H, m, H-2). Chemical shifts from δ_H 3.10 to 3.92, along with an anomeric proton resonating at [δ_H 4.42 (1H, d, J = 8.0 Hz, H-1′), δ_C 102.5], were assigned as part of a sugar moiety, which matched well with previously reported data, and subsequently confirmed as β-D-glucopyranose by HPLC (t_R = 20.2 min, Figures S66 and S67). The ¹H and ¹³C NMR data (Table 2 and Table 3) of 5 were similar to those of a known compound, (6S,7aS)-6-hydroxy-4,4,7a-trimethyl-6,7-dihydro-5H-1-benzofuran-2-one [20]. Only one more glucose unit appeared in 5, with the chemical shift value from δ_C 63.9 of (6S,7aS)-6-hydroxy-4,4,7a-trimethyl-6,7-dihydro-5H-1-benzofuran-2-one to δ_C 72.9 of 5. HMBC correlations were observed among H-9 (δ_H 1.28) with C-3, C-4, C-5, and C-10 (δ_C 48.9, 35.9, 183.8 and 30.1, respectively); H-11 (δ_H 1.59) with C-1, C-5, and C-8 (δ_C 45.6, 183.8, and 88.3); and H-6 (δ_H 5.77) to C-4, C-5, C-7, and C-8 (δ_C 35.9, 183.8, 173.8 and 88.3). Glucose was assigned at C-2 by the HMBC correlation (Figure 4) of H-1′ (δ_H 4.42) with C-2 (δ_C 72.9). The ROESY data of compound 5 indicated that H-2 was α-oriented by correlations of δ_H 4.27 (H-2) with δ_H 1.31 (CH₃-9) and δ_H 1.59 (CH₃-11). At last, the ECD analysis of 5 (2R,8R) and the enantiomer (2S,8S) were performed using the TDDFT method. The calculated ECD spectra for (2S,8S) were in agreement with the experimental spectra of 5 (Figure 5). Taking these data together, compound 5 was identified as (2S,8S)-isololiolide β-D-glucopyranoside.

Due to the similar 1D and part of 2D NMR data (HSQC, HMBC) of compounds 6a, 6b, and 5 (Table 2 and Table 3), the configuration of 6a and 6b was analyzed by ROESY experiment, and further verified by ECD calculations performed by the TDDFT method. ROESY results of stereoisomers 6a and 6b suggested that the H-2 (δ_H 4.28) of 6a was correlated with CH₃-9 (δ_H 1.27), while that of 6b was correlated with CH₃-10 (δ_H 1.45). At the same time, no correlations were found between H-2 (δ_H 4.28 and 4.30) and CH₃-11 (δ_H 1.75 and 1.74) of both compounds. In addition, the calculated ECD spectra for (2S,8R) were in agreement with the experimental data of 6a, and the same spectra for (2R,8S) with the experimental data of 6b (Figure 5). In general, compound 6a was identified as (2S,8R)-loliolide β-D-glucopyranoside and compound 6b was identified as (2R,8S)-loliolide β-D-glucopyranoside.

2.2. The Protective Effects of Compounds 1–20 on Hypoxia/Reoxygenation-Injured Neonatal Rat Cardiomyocytes

An in vitro experimental model of hypoxia/reoxygenation (H/R) injury in neonatal rat cardiomyocytes was developed to evaluate the protective effects of isolated compounds 1–20. Neonatal rat cardiomyocytes were subjected to H/R injury for four hours and subsequently evaluated for the potential protective effects of compounds 1–20. The survival of H/R-treated neonatal rat cardiomyocytes was slightly increased after treatment with compounds 2, 4, 6b, 7, 9, 17, and 20 compared to the control group (H/R treatment only), suggesting that those compounds may have a protective effect on H/R-injured neonatal rat cardiomyocytes (Figure S65).

2.3. Chemotaxonomic Significance

In the present study, 31 compounds were identified from the aerial part of A. italica (

Table 4. A comparison of the compounds from Anchusa italica with other species in the genus Anchusa or Boraginaceae ^1,2,3.

Number	Compound	Type	A. italica
1	2α,3β,19α-trihydroxy-23-formyl-urs-12-en-28,21β-olide	triterpene	1, 2
2	2α,3β,21,24-tetrahydroxyoleanan-12-en-28-oic acid	triterpene	1
3	(2R,6R,9S)-9-Hydroxy-4-megastigmen-3-one-2-O-β-D-glucopyranoside	norisoprenoids	1, 2
4	(2R,6S,9S)-9-Hydroxy-megastigman-4,7-dien-3-one-2-O-β-D-glucopyranoside	norisoprenoids	1, 2
5	(+)-Isololiolide β-D-glucopyranoside	norisoprenoids	1, 2
6a	(2S,8R)-Loliolide β-D-glucopyranoside	norisoprenoids	1, 2
6b	(2R,8S)-Loliolide β-D-glucopyranoside	norisoprenoids	1, 2
7	Niga-ichigoside F 1	triterpene	1
8	Niga-ichigoside F 2	triterpene	1
9	Pinfaensin	triterpene	1,
10	Glucosyl tormentate	triterpene	1
11	23-Hydroxytormentic acid	triterpene	1
12	24-epi-pinfaensic acid	triterpene	1
13	19α-Hydroxyasiatic acid	triterpene	1
14	(+)-Vomifoliol	norisoprenoids	1
15	Lippianoside E	norisoprenoids	1, 3
16	3-oxo-α-ionol-β-D-glucopyranoside	norisoprenoids	1, 3
17	Asysgangoside	norisoprenoids	1
18	(6S,9R)-Roseoside	norisoprenoids	1, 3	Borago officinalis L.
19	Sammangaoside B	norisoprenoids	1, 3
20	(+)-Isololiolide	norisoprenoids	1, 3
21	Tricin	flavone	1, 3
22	6-Hydroxykaempferol 3-β-rutinoside	flavonol	1, 3
23	Kaempferol 3-O-rutinoside	flavonol	1, 3	Anchusa strigosa Banks & Sol.
24	Narcissin	flavonol	1, 3	Heliotropium angiospermum
25	(+)-Syringaresinol	lignans	1, 3	Moltkia aurea Boiss.
26	(+)-Mediaresinol	lignans	1
27	Dehydrodiconiferyl alcohol	lignans	1, 3
28	Vibruresinol	lignans	1, 3
29	Dehydrodiconiferyl alcohol 4-O-β-D-glucopyranoside	lignans	1, 3
30	Syringaresinol-4′-O-β-D-glucopyranoside	lignans	1, 3	Moltkia aurea Boiss.

¹ Reported in this study; ² New compound; ³ Compound reported from Anchusa italica for the first time.

). They were categorized as triterpenes (1, 2, and 7–13), norisoprenoids (3–6 and 14–20), flavonoids (21–24), and lignans (25–30). Among these compounds, three classes of compounds including norisoprenoids, flavonoids, and lignan derivatives were isolated from A. italica, which have been reported in other Anchusa species and even some species in the Boraginaceae family (Table 4). Compounds 18, 23, 24, 25, and 30 were reported in Borago officinalis L. [7], A. strigose Banks and Sol. [12], Heliotropium angiospermum [13], and Moltkia aurea Boiss. [14], leading to a close relationship from the perspective of chemical composition. Compounds 15, 16, 18–25, and 27–30 were found for the first time in A. italica, which may complement the results from existing investigations. In previous studies, triterpenes isolated from A. italica were rarely identified; thus, compounds 1, 2, and 7–13 might expand the use of triterpenes, especially ursane triterpenoids, as chemical taxonomic markers for the classification of the genus Anchusa.

3. Materials and Methods

3.1. General Experimental Procedures

NMR spectra were obtained on a Bruker AM-500 MHz spectrometer (Bruker, Karlsruher, Germany) using TMS as an internal reference. HR-ESI-MS was performed on a quadrupole ion trap high-resolution mass spectrometer (ThermoFisher Scientific, Waltham, MA, USA). CD spectra were obtained using a Chirascan detector (Applied Photophysics Limited Shanghai Representative Office, Shanghai, China). Optical rotation values were recorded on a Rudolph Auto pol IV polarimeter (Rudolph, Hackettstown, NJ, USA). UV full-wavelength scanning was conducted on a Shimadzu UV-2401 PC spectrophotometer (Shimadzu, Kyoto, Japan). CC was performed using silica gel (200–300 mesh, 300–400 mesh, Qingdao Haiyang Chemical Group Co., Qingdao, China). Thin-layer chromatography was performed on silica gel GF254 (Qingdao Haiyang Chemical Group Co., China). D101 macroporous resin was purchased from Tianjin Yunkai Company, China. MCI gel was purchased from Mitsubishi Chemical Group Co. (Mitsubishi Chemical Group, Toyko, Japan). Semi-preparative HPLC was performed on a QingBoHua HPLC system (QingBoHua Technology, Beijing, China) equipped with a binary high-pressure pump, a dual–wavelength UV detector, and a C18 column (250 mm × 10 mm, 5 μm, YMC Co., Ltd., Kyoto, Japan). Other chemical reagents were purchased from Sinopharm Chemical Reagent Co., Ltd., Shanghai, China.

The ECD calculations were carried out using Gaussian 09. Systematic conformational analyses for compounds 3, 4, 5, 6a, and 6b were performed via Confab using the MMFF94 molecular mechanics force field calculation. All of these conformers were further optimized using time-dependent density functional theory (TDDFT) at the B3LYP/6–311G (d, p) level in methanol with the IEFPCM model. Vibrational frequency analysis confirmed the stable structures. Based on the optimized structures, the ECD calculation was conducted using TDDFT with a total of 60 excited states. The ECD spectrum was simulated in SpecDis using a Gaussian function with half–bandwidths of 0.3 eV for compounds 3, 4, and 5.

3.2. Plant Material

The whole plant of A. italica was collected from Xinjiang Uygur Autonomous Region. The plant was picked and identified as A. italica by researcher Shi Leiling of the Xinjiang Institute of Traditional Chinese Medicine and Ethnomedicine. The specimen (A20220907) is stored in the Museum of Traditional Chinese Medicine, College of Traditional Chinese Medicine, Xinjiang Medical University.

3.3. Extraction and Isolation

The powder (2.5 kg) of A. italica was extracted in 70% ethanol under reflux (3 × 2 h each time). The ethanol extract was evaporated under reduced pressure to obtain the total extract (356 g). It was dissolved in water and then extracted three times sequentially with PE, CH₂Cl₂, EtOAc, and BuOH to provide the fractions accordingly.

The ethyl acetate soluble fraction (21 g) was subjected to CC (9.4 × 40 cm) over silica gel (200–300 mesh) and eluted with CH₂Cl₂/MeOH (100:0–70:30, v/v) to yield three fractions (Y1–3). Y1 (6 g) was subjected to CC (3 × 60 cm) over silica gel (300–400 mesh) and eluted with PE-EtOAc (85:15, v/v) to yield two fractions (Y11–12). Y11 (150 mg) was prepared by HPLC using acetonitrile (MeCN)/H₂O (35:65, 210/254 nm) to provide compounds 20 (t_R = 20 min, 3 mg), 21 (t_R = 23 min, 2.6 mg), and 14 (t_R = 27 min, 2.5 mg). Y12 (200 mg) was prepared by HPLC using MeCN/H₂O (30:70, 210/254 nm) to provide compounds 15 (t_R = 14 min, 3 mg), 25 (t_R = 18 min, 13 mg), 26 (t_R = 22 min, 3 mg), and 1 (t_R = 29 min, 7 mg). Y2 (3.8 g) was subjected to CC (3 × 60 cm) over silica gel (300–400 mesh) and eluted with CH₂Cl₂/MeOH (50:1–10:1, v/v) to yield two fractions (Y21–22). Y21 (100 mg) was prepared by HPLC using MeCN/H₂O (25:75, 210/254 nm) to provide compounds 27 (t_R = 16 min, 1.8 mg) and 28 (t_R = 20 min, 7 mg). Y22 (700 mg) was prepared by HPLC using MeCN/H₂O (20:80, 210/254 nm) to provide compounds 7 (t_R = 16 min, 134 mg), 8 (t_R = 23 min, 86 mg), and 9 (t_R = 29 min, 128 mg). Y3 (350 mg) was prepared by HPLC using MeCN/H₂O (18:82, 210/254 nm) to provide compounds 10 (t_R = 20 min, 12.6 mg), 11 (t_R = 36 min, 23.4 mg), 23 (t_R = 43 min, 3.5 mg), and 24 (t_R = 45 min, 3.6 mg).

The BuOH soluble fraction (45.87 g) was subjected to CC (9.4 × 40 cm) over silica gel (200–300 mesh) and eluted with CH₂Cl₂/MeOH (100:0–60:40, v/v) to yield two fractions (Z1-Z2). Z1 (3.86 g) was subjected to CC (3 × 60 cm) over silica gel (300–400 mesh) and eluted with PE-EtOAc (75:25, v/v) to yield two fractions (Z11-Z12). Z11 (160 mg) was prepared by HPLC using MeCN/H₂O (20:80, 210/254 nm) to provide compounds 3 (t_R = 15.5 min, 2.1 mg), 16 (t_R = 32 min, 1.0 mg), and 12 (t_R = 38 min, 7.3 mg). Z12 (230 mg) was prepared by HPLC using MeCN/H₂O (17:83, 210/254 nm) to provide compounds 22 (t_R = 36 min, 4 mg), 29 (t_R = 39 min, 2.3 mg), and 30 (t_R = 45 min, 10.7 mg). Z2 (2.5 g) was subjected to CC (3×60 cm) over silica gel (300–400 mesh) and eluted with CH₂Cl₂/MeOH (70:30, v/v) to yield three fractions (Z21-Z23). Z21 (300 mg) was prepared by HPLC using MeCN/H₂O (15:85, 210/254 nm) to provide compounds 13 (t_R = 17 min, 3.4 mg), 4 (t_R = 24 min, 0.8 mg), and 2 (t_R = 29 min, 5.4 mg). Z22 (200 mg) was prepared by HPLC using MeCN/H₂O (12:88, 210/254 nm) to provide compounds 17 (t_R = 20 min, 23 mg), 18 (t_R = 24 min, 6 mg), and 19 (t_R = 36 min, 10.5 mg). Z23 (150 mg) was prepared by HPLC using MeCN/H₂O (10:90, 210/254 nm) to provide compounds 5 (t_R = 28 min, 1.3 mg), 6a (t_R = 30 min, 0.8 mg), and 6b (t_R = 32 min, 0.8 mg).

Spectroscopic Data

2α,3β,19α-trihydroxy-23-formyl-urs-12-en-28,21β-olide (1): yellow amorphous powder; ${\left[\alpha \right]}_{\mathrm{D}}^{20}$ = +70.4° (c 0.03, MeOH); UV (MeOH) λ_max 205 nm; ¹H and ¹³C NMR data (CD₃OD, 500 and 125 MHz), see Table 1; HR-ESI-MS m/z 501.3211 [M + H]⁺ (calcd for C₃₀H₄₅O₆, 501.3200). Supplementary data were available at Figures S1–S10.

(2R,6R,9S)-9-hydroxy-4-megastigmen-3-one-2-O-β-D-glucopyranoside (3): colorless oil; ${\left[\alpha \right]}_{\mathrm{D}}^{20}$ = +28.0° (c 0.05, MeOH); UV (MeOH) λ_max 240 nm; ¹H and ¹³C NMR data (CD₃OD, 500 and 125 MHz), see Table 2 and Table 3; HR-ESI-MS m/z 389.2170 [M + H]⁺ (calcd for C₁₉H₃₃O₈, 389.2160). Supplementary data were available at Figures S11–S19, S56, Tables S1–S8.

(2R,6S,9S)-9-hydroxy-megastigman-4,7-dien-3-one-2-O-β-D-glucopyranoside (4): ${\left[\alpha \right]}_{\mathrm{D}}^{20}$ = −38.1° (c 0.05, MeOH); UV (MeOH) λ_max 205 nm; ¹H and ¹³C NMR data (CD₃OD, 500 and 125 MHz), see Table 2 and Table 3; HR-ESI-MS m/z 387.2013 [M + H]⁺ (calcd for C₁₉H₃₁O₈, 387.2017). Supplementary data were available at Figures S20–S28, S57–58, Tables S9–S32.

(+)-isololiolide β-D-glucopyranoside (5): ${\left[\alpha \right]}_{\mathrm{D}}^{20}$ = +4.0° (c 0.02, MeOH); UV (MeOH) λ_max 210 nm; ¹H and ¹³C NMR data (CD₃OD, 500 and 125 MHz), see Table 2 and Table 3; HR-ESI-MS m/z 381.1519 [M + Na]⁺ (calcd for C₁₇H₂₆O₈Na, 381.1522). Supplementary data were available at Figures S29–S37, Tables S33–S34.

(2S,8R)-loliolide β-D-glucopyranoside (6a): ${\left[\alpha \right]}_{\mathrm{D}}^{20}$ = −51.4° (c 0.05, MeOH); UV (MeOH) λ_max 210 nm; ¹H and ¹³C NMR data (CD₃OD, 500 and 125 MHz), see Table 2 and Table 3; HR-ESI-MS m/z 381.1518 [M + Na]⁺ (calcd for C₁₇H₂₆O₈Na, 381.1522). Supplementary data were available at Figures S38–S46, Tables S35–S36.

(2R,8S)-loliolide β-D-glucopyranoside (6b): ${\left[\alpha \right]}_{\mathrm{D}}^{20}$ = +5.7° (c 0.03, MeOH); UV (MeOH) λ_max 210 nm; ¹H and ¹³C NMR data (CD₃OD, 500 and 125 MHz), see Table 2 and Table 3; HR-ESI-MS m/z 381.1518 [M + H]⁺ (calcd for C₁₇H₂₆O₈Na, 381.1522). Supplementary data were available at Figures S47–S55, Tables S35–S36.

3.4. Acid Hydrolysis of New Compounds

The new compounds 3, 4, 5, 6a, and 6b (1 mg each) were dissolved in 2 mL of CF₃COOH (4 mol/L) solution and reacted at 95 °C for 3 h. The reaction was cooled to room temperature, the reaction solution was extracted with an equal volume (2 mL) of CH₂Cl₂ three times, and the aqueous layer after extraction was completely concentrated. A 1 mg amount of D-glucose (China Academy of Food and Drug Administration, Lot No. 110833) and the extracted aqueous phase were taken into four 25 mL round-bottomed flasks, and anhydrous pyridine (0.5 mL) and L-cysteine methyl ester hydrochloride (1.0 mg) were added, respectively; the reaction was carried out at 60 °C for 1 h. After cooling, 5 μL of o-toluene isothiocyanate was added to each reaction, and the reaction was carried out at 60 °C for 1 h. The solution was cooled to room temperature, the reaction solutions of the above four groups were extracted with an equal volume (2 mL) of CH₂Cl₂ three times, and the reaction solution was completely concentrated. The reaction solution of the above four groups was diluted 1-fold with pyridine, the samples (5 μL) were analyzed by HPLC (MeCN/H₂O 25:75, 0.8 mL/min) with a detection wavelength of 250 nm, and the analytical column was a Waters chromatographic column (4.6 × 250 mm, 5 μm particle size). The acid hydrolysis products of compounds 3, 4, 5, 6a, and 6b showed peaks at t_R 20.2 min; D-glucose showed peaks at t_R 20.2 min (Supporting Information Figures S59–S64).

3.5. Quantum Chemical Calculations (Computational NMR and ECD)

The theoretical calculations were carried out using Gaussian 09. At first, all conformers were optimized at PM6. Room-temperature equilibrium populations were calculated according to the Boltzmann distribution law, based on which dominative conformers of the population over 1% were kept. The chosen conformers were further optimized at B3LYP/6-31G (d, p) in gas phase. Vibrational frequency analysis confirmed the stable structures. NMR calculations were carried out using the Gauge-Including Atomic Orbitals (GIAO) method at mPW1PW91/6-311 + G (2d, p) level in methanol simulated by the IEFPCM model. The TMS-corrected NMR chemical shift values were averaged according to the Boltzmann distribution and fitted to the experimental values by linear regression. The calculated ¹³C- and ¹H-NMR chemical shift values of TMS in methanol were 49.2 and 4.9 ppm respectively. To confirm the conclusions of NMR calculations, DP4+ analysis was also performed.

$${\displaystyle \frac{N_i}{N}}={\displaystyle \frac{g_i{e}^{-\frac{E_i}{k_{\mathrm{B}}T}}}{{\displaystyle \sum g_i{e}^{-\frac{E_i}{k_{\mathrm{B}}T}}}}}$$

where N is the number of conformer i with energy and degeneracy at temperature T, and k_B is Boltzmann constant.

ECD calculations were conducted at a B3LYP/6-311G (d, p) level in methanol with the IEFPCM model using time-dependent density functional theory (TDDFT). The rotatory strengths for 30 excited states were calculated. The ECD spectrum was simulated using the ECD/UV analysis tool on the Yinfo Cloud Platform (https://cloud.yinfotek.com/ (27 April 2025)) by overlapping Gaussian functions for each transition, where σ represents the width of the band at a height of 1/e, while ΔE_i and R_i are the excitation energies and rotatory strengths for transition i, respectively. The spectra of the enantiomers were produced directly through mirror inversion about the horizontal axis.

$$\Delta \epsilon \left(E\right)={\displaystyle \frac{1}{2.241\times {10}^{-39}}}\times {\displaystyle \frac{1}{\sqrt{2\pi \sigma }}}\sum _i^A\mathsf{\Delta }{E}_i{R}_i{e}^{-{\left({\displaystyle \frac{E-E_i}{2\sigma }}\right)}^2}$$

3.6. Cell Culture and Hypoxia/Reoxygenation

Primary cultured neonatal rat cardiomyocytes were cultured at 37 °C with 95% air and 5% CO₂ in DMEM supplemented with 10% FBS, 100 μg/mL streptomycin, and 100 units/mL penicillin. When the cells had grown to 70–80% confluence, they were pretreated with compounds or vehicle (DMSO) for 4 h (a time point with better efficacy in cardiomyocyte protection). The cells were then exposed to an anaerobic medium (serum- and glucose-free) in an hypoxia incubator chamber (SANYO, IX51) with an anoxic mixture gas (95% N₂ and 5% CO₂) for 6 h at 37 °C followed by reoxygenation for 4 h with fresh culture medium (95% air and 5% CO₂) to simulate H/R injury in isolated hearts.

4. Conclusions

The phytochemical investigation of ethanolic extracts of A. italica led to the isolation and characterization of 31 compounds, including a new ursane triterpene (1) and five new norisoprenoids (3, 4, 5, 6a, and 6b). The chemical structures of the new compounds (including absolute configurations) were fully confirmed through comprehensive analyses involving 1D and 2D NMR, HR-ESI-MS, acid hydrolysis, and computational methods for ECD calculation. The protective effects of compounds 1–20 against hypoxia/reoxygenation (H/R)-induced cardiomyocyte injury were tested. The chemical taxonomic significance of the triterpenes (1, 2, and 7–13) may expand the use of terpenoids as a chemotaxonomy marker for the classification of the genus Anchusa.