Simultaneous Determination of 78 Compounds of Rhodiola rosea Extract by Supercritical CO2-Extraction and HPLC-ESI-MS/MS Spectrometry

The plant Rhodiola rosea L. of family Crassulaceae was extracted using the supercritical CO2-extraction method. Several experimental conditions were investigated in the pressure range of 200–500 bar, with the used volume of cosolvent ethanol in the amount of 1% in the liquid phase at a temperature in the range of 31–70°C. The most effective extraction conditions are pressure 350 bar and temperature 60°C. The extracts were analyzed by HPLC with MS/MS identification. 78 target analytes were isolated from Rhodiola rosea (Russia) using a series of column chromatography and mass spectrometry experiments. The results of the analysis showed a spectrum of the main active ingredients Rh. rosea: salidroside, rhodiolosides (B and C), rhodiosin, luteolin, catechin, quercetin, quercitrin, herbacetin, sacranoside A, vimalin, and others. In addition to the reported metabolites, 29 metabolites were newly annotated in Rh. rosea. There were flavonols: dihydroquercetin, acacetin, mearnsetin, and taxifolin-O-pentoside; flavones: apigenin-O-hexoside derivative, tricetin trimethyl ether 7-O-hexosyl-hexoside, tricin 7-O-glucoronyl-O-hexoside, tricin O-pentoside, and tricin-O-dihexoside; flavanones: eriodictyol-7-O-glucoside; flavan-3-ols: gallocatechin, hydroxycinnamic acid caffeoylmalic acid, and di-O-caffeoylquinic acid; coumarins: esculetin; esculin: fraxin; and lignans: hinokinin, pinoresinol, L-ascorbic acid, glucaric acid, palmitic acid, and linolenic acid. The results of supercritical CO2-extraction from roots and rhizomes of Rh. rosea, in particular, indicate that the extract contained all biologically active components of the plant, as well as inert mixtures of extracted compositions.


Introduction
e plant Rhodiola rosea L. of family Crassulaceae is widely used in traditional medicine and traditional medical systems (Tibetan, Chinese, and Korean). Rhizomes and plant roots are mainly used for the preparation of medicinal products [1,2]. e plant has an established popular name "golden root." e name is determined not only by the color of the rhizome but also by its high price. e main medicinal raw adaptogens such as Panax ginseng and Eleutherococcus. Adaptogens are a pharmacological group of drugs of natural or synthetic origin, which can increase the body's resistance to various adverse environmental conditions [3][4][5].
Information on the content of salidroside and rosavin in Rh. rosea is numerous and contradictory [10,11]; Zang et al., 2019). Researchers still have not come to a consensus on the localization and activity of specialized biosyntheses, the nature of seasonal changes in glycoside content, and the variability in the accumulation of these substances in wild and cultivated plants [12][13][14].
Detailed comparative studies of the content of salidroside and rosavin in the organs of wild-growing and cultivated plants were carried out. Performed using a unified determination method showed the presence of glycosides only in the roots and caudex. e presence of rosavin and salidroside in the aerial organs (stems, leaves, inflorescences, and seeds) was not detected in any case [15].
Plants from different places of growth differed significantly in the accumulation of individual glycosides. e content of salidroside in the plant caudex varied from 9 to 20 mg/g dry weight. e largest accumulation of this glycoside was characterized by plants growing on rocks on the coast of the Barents Sea (Norway), as well as Ural plants growing on outcrops of bedrock with an insignificant soil layer. e minimum salidroside content was found in Altai plants. e highest content of rosavin (32 mg/g) was found in the caudex of plants of the subalpine ecotope in the Polar Urals, the lowest (10-12 mg/g) being in plants growing on the islands and the coast of the Barents Sea. Cultivated plants were not inferior for accumulation of rosavin to wild plants.
Differences in the accumulation of glycosides by plants of various ecotopes were revealed. So, in the Subpolar Urals, in the caudex of plants growing in faults and on ledges of rocks, more salidroside accumulates, but these plants were characterized by a low content of rosavin, 1.5-2 times less than in plants of the subalpine ecotope [15].
Cinnamic glycosides, and in particular rosavin, are believed to be the hallmark of the chemotaxonomic trait of Rh. rosea [16,17] [15]. In official medical practice, Rh. rosea root extract is intended for oral administration as a tonic and immunomodulating therapeutic agent. In the study of alcoholic extracts of Rh. rosea, their hepatoprotective, nootropic, cardioprotective, and antiarrhythmic properties were clearly demonstrated [18][19][20].
Cinnamic glycosides, also called cinnamyl glycosides and salidroside, are the main carriers of the biological activity of Rh. rosea, causing a positive pharmacological effect. With the presence of rosavin, rosin, and rosarin, many researchers attribute the increased biological activity of extracts of Rh. rosea, compared with drugs from other species of Rhodiola. Studies have shown the stimulating effect of drugs on the central nervous system. Of great interest is the ability of Rh. rosea to increase the body's resistance to the effects of various stress factors [21,22]. Rh. rosea extract has immune stimulating, hepatoprotective, and antimicrobial effects [23,24]. Studies have also been conducted on the antitumor effect of Rh. rosea extract [25][26][27].
is study considers the effectiveness of supercritical CO 2 -extraction of biologically active substances from roots and rhizomes of Rh. rosea. Previously, the authors of this article successfully used supercritical CO 2 extraction to obtain biologically active substances from plants of the Far Eastern taiga Panax ginseng, Rhododendron adamsii, Schisandra chinensis, and sea cucumber which are extremely popular in traditional medicine of Southeast Asia [28,29].
Supercritical fluid extraction (SFE) has been used since 1960s to analyze food and pharmaceutical products, isolate biologically active substances, and determine lipid levels in food and levels of toxic substances. In addition, the products do not have residues of organic solvents, which occur with conventional extraction methods, and solvents can be toxic, for example, in the case of methanol and n-hexane. High selectivity, easy solvent removal from the final product, and the use of moderate temperatures in the extraction process are the main attractive factors of SFE, leading to a significant increase in research for use in the food and pharmaceutical sectors [30,31].
In Sweden, an article was published in 2009 that examined the extraction of rosavin from the roots and rhizomes of Rh. rosea using supercritical CO 2 -extraction. In this case, water was selected as a modifier of supercritical extraction, which gave a synergistic effect on the extraction yield of rosavin [32]. In China, researchers used supercritical CO 2 -extraction with ethanol modifier [33]. e purpose of this study was to extract the maximum amount of salidroside from the roots of Rh. rosea. e extraction conditions were chosen so that the yield of salidroside during supercritical extraction was much higher than the yield of the product when using classical extraction using a Soxhlet apparatus.

2
Biochemistry Research International e results of SC-CO 2 -extraction of from roots and rhizomes of Rh. rosea, in particular, indicate that when using this technology, the extract contained all biologically active components of the plant, as well as inert mixtures of extracted compositions.

Experimental
2.1. Materials. Ground, dried root of Rh. rosea was obtained from the area near Lake Baikal, Russia. All samples were morphologically authenticated according to the current standard of Russian Pharmacopeia [34]. e volume weighted mean diameter of the powder was found as 550 μm, as determined by dynamic light scattering (Hydro 2000MU Malvern Instruments Ltd.).

Chemicals and Reagents.
HPLC-grade acetonitrile was purchased from Fisher Scientific (Southborough, UK), and MS-grade formic acid was purchased from Sigma-Aldrich (Steinheim, Germany). Ultrapure water was prepared from Siemens Ultra-Clear water purification system (Siemens Water Technologies, Germany), and all other chemicals were analytical grade.

Supercritical Fluid Extraction.
A supercritical fluid extraction system was ar SFE-500F-2-FMC50 ( ar Technology Inc., Pittsburgh, PA, USA) which is used in supercritical extraction. CO 2 was compressed to the required pressure using a supercritical extraction compressor ( ar SFC, USA). A hot casing string heated the extraction vessel; the temperature was regulated by a thermostat (±1°C). A Biochemistry Research International 3 metering valve controlled the pressure. Shredded Rhodiola roots (50 g) were wrapped in a filter paper, charged to a oneliter extractor, and extracted with supercritical CO 2 compressed to a supercritical state at a liquid flow rate of 250 g/ min. Seven SFE extracts were obtained under different pressure conditions (100-400 bar) and temperatures (31-70°C). Ethanol served as the cosolvent in all cases. e extracts were collected in a separator. e pressure and temperature of the supercritical CO 2 were optimized experimentally to achieve the maximum yield of the product during extraction.

Mass Spectrometry. MS analysis was performed on an ion trap amaZon SL (Bruker Daltoniks, Germany) equipped
with an ESI source in the negative ion mode. e optimized parameters were obtained as follows: ionization source temperature, 70°C; gas flow, 4l/min; nebulizer gas (atomizer), 7.3 psi; capillary voltage, 4500 V; end plate bend voltage, 1500 V; fragmentary, 280 V; and collision energy, 60 eV. An ion trap was used in the scan range m/z 100-1.700 for MS and MS/MS. e capture rate was one spectrum/s for MS and two spectra/s for MS/MS. Data collection was controlled by Windows software for Bruker Daltoniks. All experiments were repeated three times. A two-stage ion separation mode (MS/MS mode) was implemented.

Results and Discussion
Several experimental conditions were investigated in the pressure range 200-500 bar, with the used volume of cosolvent ethanol in the amount of 1% in the liquid phase at a temperature ranging 31-70°C. Ethanol was used as the modifier due to its high solubility in CO 2 and high polarity and ability to disturb solute-plant matrix bonding. As a result of using a wide range of pressures and temperatures empirically, the most efficient extraction conditions were found for extracting target analytes from the Rh. rosea roots. e most effective extraction conditions are pressure 350 bar and temperature 60°C ( Figure 1).
Obtaining chemical profiles is an extremely important result in the biological analysis system. In this work, we used the HPLC-ESI-MS/MS method with additional ionization and analysis of fragmented ions. High accuracy mass spectrometric data were recorded on an ion trap amaZon SL (Bruker Daltoniks) equipped with an ESI source in the negative ion mode. e two-stage ion separation mode (MS/ MS mode) was implemented. Figure 2 shows the distribution density of the analyzed chemical profiles in the ion chromatogram of the Rh. rosea supercritical CO 2 -extract, realized by mass spectrometry in the two-stage ion separation mode (MS/MS mode).
A unifying system table consists of the molecular masses of the target analytes isolated from the supercritical CO 2extract of Rh. rosea for ease of identification (Table 2). e CID spectrum (collision induced dissociation spectrum) in negative ion modes of Rhodioloside B from Rh. rosea is shown in Figure 3. e [M−H] − ion produced two fragments with m/z 447.00 and m/z 219.49 ( Figure 3). e fragment ion with m/z 447.00 yields a daughter ion at m/z 314.98. e interpretation of the observed MS/MS spectra in comparison with those found in the literature was the main tool for putative identification of polyphenols. It was identified in the bibliography in extracts from Rh. rosea [50], from Rhodiola crenulata [35]. e CID spectrum in the negative ion mode of luteolin-7-O-α-L-rhamnoside from Rh. rosea is shown in It was identified in the bibliography in extracts from Rhodiola crenulata [35]. e CID spectrum in the positive ion mode of catechin from Rh. rosea is shown in Figure 5. e [M+H] + ion produced fragments with m/z 273.14 and m/z 217.09 ( Figure 5). It was identified in the bibliography in extracts from Rh. rosea [50], from strawberry, cherimoya [36], and pear [45].
We isolated 78 target analytes from Rhodiola rosea L. (Crassulaceae) using a series of column chromatography and mass spectrometry experiments. e structures were elucidated using the data of stepwise fragmentation of ions during MS/MS spectrometry and compared with spectroscopic data in the literature. It is accepted that glycosides of cinnamon alcohol, and in particular Rosavin, are a distinctive chemotaxonomic sign of Rh. rosea [17]. However, lately, information has appeared in the literature on the presence of this glycoside in other species of the genus Rhodiola L. [15]. us, we can summarize the research that the supercritical extraction of the roots of Rh.

Conclusions
e Rhodiola rosea L. family Crassulaceae contains a large number of polyphenolic compounds and other biologically active substances. In this work, we tried to conduct a comparative metabolomic study of biologically active substances of Rh. rosea obtained from the area near Lake Baikal, Russia. HPLC in combination with a Bruker Daltoniks ion trap (tandem mass spectrometry) was used to identify target analytes in extracts.
e results showed the presence of 78 polyphenols and other compounds corresponding to the Rhodiola rosea family Crassulaceae L. species. In addition to the reported metabolites, 29 metabolites were newly annotated in Rh. e findings may support future research into the production of various pharmaceutical and dietary supplements containing Rh. rosea extracts. A wide variety of biologically active compounds opens up rich opportunities for the creation of new drugs and biologically active additives based on extracts from family Crassulaceae.

Data Availability
No data were used to support this study.

Conflicts of Interest
e authors declare that they have no conflicts of interest.