Study on the Alkaloids in Tibetan Medicine Aconitum pendulum Busch by HPLC–MSⁿ Combined with Column Chromatography

Abstract

A rapid, convenient and effective identification method of alkaloids was established and an attempt on isolating and analyzing the alkaloids in Aconitum pendulum Busch was conducted successfully. In this article, four high-content components including deoxyaconitine, benzoylaconine, aconine and neoline were isolated by using column chromatography. HPLC–MSⁿ was employed to deduce the regulations of fragmentation of diterpenoid alkaloids which displayed a characteristic behavior of loss of CO(28u), CH₃COOH(60u), CH₃OH(32u), H₂O(18u) and C₆H₅COOH(122u). Then, according to fragmentation regulation of mass spectrometry, 42 alkaloids were found in A. pendulum. Among them, 38 compounds were identified and 29 alkaloids were reported for the first time for this herb. Therefore, this means that HPLC–MSⁿ combined with column chromatography could work as an effective and reliable tool for rapid identification of the chemical components of herbal medicine.

Introduction

Aconitum pendulum Busch (A. pendulum) is the dried root of Aconitum flavum Hand. Mazz (1), which is widely distributed, mainly in Tibet Autonomous Region, Yunnan province, Qinghai province and Sichuan province, China (2–5). From a long time ago, A. pendulum has been used as an ethnomedicine among Qiang, Tibetan, Hui and other ethnic groups (2). It's a powerful herb with multifunction, such as treatment of influenza, fever caused by varieties of infectious diseases, rheumatoid arthritis and joint pain, although toxicity results easily from overdose (2, 6, 7).

Diterpenoid alkaloids were considered the principal active component of A. pendulum (8–10). However, so far, the constituents in A. pendulum and the material basis of its efficacy are still unclear. Hence, further research studies on the composition of A. pendulum are extremely necessary. The traditional extraction and separation technologies of compounds have general suitability. Nevertheless, with these it is difficult to isolate the minor and trace constituents (11). In recent years, HPLC–MSⁿ, a rapid and convenient means, has started to play a significant role in the identification of complex components, which has made up for the shortage of traditional methods (11–13). This method holds the following obvious advantages: almost all of the compounds can be detected with a wide adaptability detector; difficulties with the analysis of thermally unstable compounds are easily solved; has high capabilities of separation and compound identification, and so on. In addition, all structure information and molecular weight of each component can be provided for the identification of its structure (14–17). Moreover, much of the research studies in isolating and identifying alkaloids in Aconitum L. had determined that HPLC–MSⁿ can be applied to isolate and identify alkaloids (13, 18–20).

In this article, the high-content alkaloids were isolated by column chromatography first. Then, with HPLC–MSⁿ, a regular mass spectrometric pattern was rapidly generalized via analyzing the trace elements and the total ion chromatogram (TIC) for components in herbal material was obtained. Finally, based on the known fragmentation patterns and literature data, the identification of diterpenoid alkaloids was completed for the first time. This approach, traditional column chromatography combined with HPLC–MSⁿ, has a potential for rapid isolation and identification of the ingredients in different kinds of herbs simultaneously, which is a shortcut in the study of material foundation.

Experimental

Instrument and analytical conditions

HPLC–MSⁿ analysis was performed on an Agilent series 1100 HPLC instrument (Agilent, Waldbronn, Germany) coupled with an Agilent 1100 MSD XCT/plus ion-trap mass spectrometer (Waldbronn, Germany) via an electrospray ionization (ESI) interface. The HPLC instrument was equipped with a quaternary pump, a diode-array detector, an autosampler and a column compartment. Samples were separated on an Agilent Extend-C₁₈ column (150×4.6 mm I.D., 5 μm) equipped with a Waters Symmetry C₁₈ guard column (20 × 3.9 mm I.D., 5 μm).

Materials

Herbs of Aconitum flavum Hand. Mazz were collected from the district of Maqin, Qinghai in September 2013, and authenticated by Prof. Peng Tan. Samples were stored at the Chinese Medicine faculty of Beijing University of Chinese Medicine. HPLC-grade MeCN was purchased from Fisher (USA). Water for HPLC was double distilled and filtered with a 0.45 μm filter membrane.

The preparation of monomeric chemical substances on column chromatography

The air-dried powder of A. pendulum was extracted three times (each for 1 h) with 80% ethanol. The filtrate was concentrated under vacuum to obtain an ethanol extract (220 g). The liquid extract was dissolved in hydrochloric acid (pH = 1) as fully as possible and the pH was adjusted to 10 with concentrated ammonia. The total alkaloids were extracted with H₂O saturated n-butanol, and 23 g of extract was obtained after the solvent was recovered under vacuum. The n-butanol extract was separated by silica gel column chromatography using a gradient elution of dichloromethane–methanol (50 : 1–1 : 1). The constituents were combined into seven fractions (I–VII) after thin-layer chromatography (TLC) analysis. Through further separation and purification using silica gel column chromatography and Sephadex LH-20 as well as a method of preparative TLC, we obtained both 1 (10 mg) and 2 (15 mg) from fraction II, 3 (13 mg) from fraction III, and 4 (22 mg) from fraction V.

Sample preparation for HPLC–MSⁿ

One gram of powdered A. pendulum was ultrasonically extracted in 25 mL of methanol for 1.0 h at room temperature. The sample solution was filtered through a 0.45-μm filter membrane.

Approximately 1 mg of the four monomeric compounds isolated from A. pendulum were separately dissolved in 100 mL of methanol. Then, through a 0.45-μm filter membrane, the standard solution was obtained.

LC–MS-MS conditions

The mobile phase consisted of water containing ammonia (A) (99.6 : 0.4, v/v) and MeCN (B). A gradient program was used as follows: 33% (B) in 0–15 min, 33–50% (B) in 15–45 min, 50–80% (B) in 45–65 min. A 15-min post-run time was set to fully equilibrate the column. The flow rate was 1 mL/min. The column temperature was 30°C. The sample injection volume was 10 μL. The HPLC eluent was introduced into the ESI source of the mass spectrometer in a post-column splitting ratio of 4 : 1. For MS detection, high-purity nitrogen (N₂) was used as the nebulizing gas, and ultra-high pure helium as the collision gas. Positive-ion polarity modes were used for compound ionization. The ESI source parameters were optimized by injecting a 6 μL/min flow of neoline to obtain maximum intensities of ions. The optimized parameters in the positive-ion mode were as follows: source voltage, 4.5 kV; sheath gas (N₂), 40 arbitrary units; auxiliary gas (N₂), 10 units; capillary temperature, 330°C; capillary voltage, −30 V; and tube lens offset voltage, −20 V. In the positive ESI ion mode, the capillary voltage was 15 V and the tube lens offset voltage was 30 V. For full scan MS analysis, spectra were recorded in the range of m/z 100–1,000. The data-dependent program was set so that the two most abundant ions in each scan were selected and subjected to tandem mass spectrometry (MSⁿ, n = 4). The isolation width of precursor ions was 2.0 Th.

Results

Identification of alkaloids on column chromatography

Through TLC and the comparison of characteristic fragments, compounds 1, 2, 3 and 4 were identified as deoxyaconitine, benzoylaconine, aconine and neoline. MS^1–4 of deoxyaconitine (Figure 1), benzoylaconine (Figure 2), aconine (Figure 3) and neoline (Figure 4) were obtained.

Figure 1.

MS^1–4 spectra of deoxyaconitine.

Figure 2.

MS^1–4 spectra of benzoylaconine.

Figure 3.

MS^1–4 spectra of aconine.

Figure 4.

MS^1–4 spectra of neoline.

Establishment of MS fragmentation mechanism

According to Figures 1–4 and the previously reported fragmentation pathways [], regulations were represented as follows: peaks of [M+H]⁺ are prone to appear successively or simultaneously with the neutral loss of CH₃COOH (60 Da), CH₃OH (32 Da), H₂O (18 Da) and PhCOOH (122 Da).

Rapid analysis of the compounds in A. pendulum

In this article, the TIC of A. pendulum is shown in Figure 5, and 38 alkaloids were tentatively identified by detailed study of their fragmentation regulations (Table I).

Figure 5.

The TIC of A. pendulum.

Table I.

Compounds Identified in A. pendulum by HPLC–MS in Positive-Ion Mode

No.	Retention time	MS	MS2	MS3	MS4	Identification and type
1	1.8	656.9	482.7	422.6	372.2	Unknown
2	1.9	760.9	586.8	526.6	494.3	8-Shikimoyl-benzoylaconine (LDA)
3	2.7	704.7	586.5	526.4	494.3	8-scn-Benzoylaconine (LDA) (21)
4	2.8	344.3	298.2	266.3	237.1	Bullatine A (other) (22)
5	3.1	425.3	406.5	388.5	356.1	Senbusine A/B (ADA) (23)
6	3.2	454.6	422.5	404.4	372.2	Fuzlinea (ADA) (24)
7	3.5	456.9	424.6	392.4	360.1	Unknown
8	3.9	500.7	450.9	418.5	386.2	Aconinea (MDA) (25)
9	4.2	466.8	448.5	430.3		14-O-acetylvirescenine (MDA) (26)
10	4.5	542.6	482.7	422.8	390.3	14-Benzoylnoeline (MDA) (27)
11	6	496.7	478.6	446.6	358.3	14-O-acetyl-10-hydroxyneoline (28)
12	6.7	438.5	420.6	388.5	356.3	Neoline (ADA)a (29)
13	7	542.8	450.8	418.5	386.2	Isomer of 14-benzoylnoeline
14	7.9	376.7	358.4	340.3	322.5	Weisaconitine C (C18-) (30)
15	9	496.7	436.6	376.5	358.2	Unknown
16	10	586.8	554.7	522.6	490.2	Dehydrated 14-benzoylaconitine (MDA) (18)
17	10.5	438.5	420.6	388.6	356.3	Isomer of neoline
18	11.7	480.9	462.7	430.5	398.2	14-Acetylneoline (ADA) (31)
19	12.2	618.7	558.7	476.5	444.2	N-diethyl-aconitine (DDA) (32)
20	12.7	618.7	558.9	498.5	448.2	O-Didemethyl-aconitine (DDA) (32)
21	15	604.4	586.7	526.6	494.2	Isomer of benzoylaconine
22	16.9	360.9	342.5	324.4	237.1	12-Epinapellinea (C20-) (31)
23	18.7	604.6	554.7	522.5	490.2	Benzoylaconine (MDA) (29)
24	20.1	484.8	452.5	420.5	388.2	Deoxyaconinea (ADA) (29)
25	21.3	632.7	572.6	512.5	480.3	10-OH-hypaconitine (DDA) (29)
26	23.3	588.7	528.5	496.4	464.3	14-Benzoyldeoxyaconine (MDA) (25)
27	23.8	632.9	572.7	540.5	522.3	Mesaconitine (DDA) (33)
28	26.5	602.7	542.6	482.6	446.2	Dehydrated 10-OH-aconitine (DDA) (29)
29	27.4	662.8	602.7	542.6	510.2	10-OH-aconitine (DDA) (25)
30	28.2	526.7	466.6	406.6	374.3	Unknown
31	29.5	660.6	600.9	540.5	508.3	Yunaconitine (DDA) (34)
32	32.8	618.8	568.7	536.6	504.2	8-Methoxyl-14-benzoylaconine (MDA) (32)
33	33.2	628.6	568.5	536.4	504.2	Foresaconitine (DDA) (31)
34	34.5	586.8	536.6	504.5	472.2	Dehydrated aconitine (DDA) (18)
35	35	602.9	542.6	510.4	478.3	16-O-demethylhypaconitine (DDA) (19)
36	37.5	646.8	554.7	522.6	490.2	Isomer of aconitine
37	38.3	586.9	526.6	494.6	462.2	Isomer of dehydrated aconitine
38	41.7	630.9	570.7	520.6	488.2	Indaconitine (DDA) (35)
39	42.9	646.9	586.8	526.6	494.3	Aconitinea (DDA) (29)
40	49.3	630.7	570.8	510.6	478.3	Deoxyaconitinea (DDA) (25)
41	54.5	616.7	556.6	524.5	496.3	Hypaconitinea (DDA) (29)
42	60	688.7	628.6	568.6	508.3	3-Acetylaconitinea (DDA) (24)

^aCompounds have been reported in A. pendulum. C₁₈- refers to C₁₈-diterpenoid-type alkaloids and C₂₀- refers to C₂₀-diterpenoid-type alkaloids, respectively.

Aconitum alkaloids often share a C₁₉- or C₂₀-diterpenoid skeleton. Based on the substitute groups at the C₈ and C₁₄ positions, the C₁₉-diterpenoid-type alkaloids were commonly divided into diester-diterpenoid alkaloids (DDAs), monoester-diterpenoid alkaloids (MDAs), amine-diterpenoid alkaloids (ADAs) and lipo-diterpenoid alkaloids (LDAs). In A. pendulum, 15 DDAs were identified and the pseudo-molecular ions tend to lose 60 Da (CH₃COOH) firstly by the elimination of a molecule of AcOH at the C₈ site, which could serve as the diagnostic ion for distinguishing from the other types. The further fragmentation mainly led to the combined neutral loss of 60 Da (CO + CH₃OH) or CH₃OH in the MS³, MS⁴. There are seven MDAs that were found and common fragment pathways represented losses of CH₃OH and H₂O from MS to MS² simultaneously, which could be taken as diagnostic fragment ions for MDAs. The fragmentation behaviors of ADAs were similar to MDAs. However, ADAs were prone to lose H₂O due to their structure characteristics.

In addition, there were two LDAs that were identified. In general, the MSⁿ spectra of LDA were identical with that of the corresponding DDA, of which the C₈ sites were occupied by acetyl groups. For compound 2 and 3, the [M ]⁺ ion was at m/z 704 and 760, respectively. The prominent fragment ion at m/z 586 indicated it was an 8-lipo-benzoylaconine derivative. The former had been reported and it was 8-scn-benzoylaconine; the latter produced neutral loss of 174 Da corresponding to shikimic acid residue. Therefore, it was plausibly identified as 8-shikimoyl-benzoylaconine, which was reported for the first time.

Besides C₁₉-diterpenoid-type, there are one C₁₈-diterpenoid-type alkaloid and one C₂₀-diterpenoid-type alkaloid that were identified according to the reference materials. The pseudo-molecular ion of Weisaconitine C is assigned to 376. Its fragmentation ions were at m/z 376, 358, 340, 322; 358 was loss of H₂O from 376, 340 was loss of H₂O from 358, and 322 was loss of H₂O from 340, respectively. The pseudo-molecular ion of 12-epinapelline is assigned to 360. Its fragmentation ions were at m/z 360, 342, 324; 342 was loss of H₂O from 360, and 324 was loss of H₂O from 342.

Discussion

Selection of extraction methods

In the separation of alkaloids, the acid-dissolution and alkali-precipitation method was used, which enables enriching of alkaloids with higher concentration. However, sample preparations for HPLC–MSⁿ were extracted with methanol by means of ultrasonication in case relatively low levels of alkaloids could not be obtained.

Identification of alkaloids in A. pendulum by LC–MSⁿ

An Agilent Extend-C₁₈ column was used for reversed-phase HPLC in our work, and ammonia was added in the mobile phase to depress the tailing of the peaks of aconitum alkaloids. A gradient mobile phase was used to guarantee good efficiency of separation because of the complexity of the chemical constituents.

Four high-content components were isolated. They were regarded as representatives of diterpenoid alkaloids in A. pendulum to optimize ESI conditions and summarize the common proposed fragmentation pathways with the combination of previous studies. The result showed that the positive-ion mode is more appropriate than the negative ion mode for the identification of alkaloids. Most compounds identified in this work were isolated or identified from the genus Aconitum L.

Although the number of alkaloids is large in Aconitum species, most share a common skeleton. In A. pendulum, C₁₉-norditerpenoid skeleton is the main type and the different substituent groups and substituent positions lead to different alkaloids, which make the occurrence of isomers common. For example, the pseudo-molecule [M+H]⁺ ions in the EIC for m/z 618 (Figure 6) at 12.2 min (compound18), 12.7 min (compound 19) and 32.8 min (compound 32) indicated that they were isomers. Among them, there was a significant difference in fragment ions between compound 18 and compound 19 in MS³, and the fragment ions of compound 32 were completely different compared with those of compounds 18 and 19. According to the reference materials, they were identified as N-diethyl-aconitine, O-didemethyl-aconitine and 8-methoxyl-14-benzoylaconine.

Figure 6.

The pseudo-molecule [M+H]⁺ ions of N-diethyl-aconitine, O-didemethyl-aconitine and 8-methoxyl-14-benzoylaconine.

Conclusion

Alkaloids, the predominant components in Aconitum species, have been widely studied for their complex structures, thought-provoking chemistry and noteworthy bioactivities. To make further research studies of A. pendulum, a rapid isolation and identification method is indispensable. In this article, a method of HPLC–MSⁿ combined with column chromatography was established to identify diterpenoid alkaloids rapidly and conveniently. Compared with using column chromatography only, this method is environmental friendly, time saving and much less laborious and, at the same time, it is more accurate than only using HPLC–MSⁿ to identify compounds in complex plant extracts. Along with its successful application to A. pendulum Busch, this method could be extended to other plants and lay a foundation for herbal medicine quality evaluation.

Funding

This study was supported by National Natural Science Foundation of China (No. 30901959, 81102807).

Conflict of interest statement. None declared.