Phytochem. Anal. 10, 22–25, (1999)


High Performance Liquid Chromatographic

Quantification of Salvinorin A from Tissues of

Salvia divinorum Epling & Játiva-M


John W. Gruber, Daniel J. Siebert, Ara H. Der Marderosian and Rick S. Hock


A reversed‑phase high performance liquid chromatographic method for the determination of salvinorin A, a psychotropic diterpene isolated from the Mexican sage Salvia divinorum, has been developed. Extracts from several plant collections were examined on a C‑18 column with UV detection and isocratic elution with acetonitrile: water (45:55). This assay allowed quantification of salvinorin A in extracts of leaves and stems of S. divinorum and has also been applied to the screening of related species for the production of salvinorin A. Levels of salvinorin A in leaves range from 0.89 to 3.70 mg/g dry weight. © 1999 John Wiley & Sons, Ltd.


Keywords: Hallucinogen; high performance liquid chromatography; Mexican sage; psychotropic diterpene; Salvia divinorum; salvinorin A.




Salvia divinorum Epling et Jdtiva‑M. (Lamiaceae) is native to Oaxaca in Central Mexico. Among the indigenous Mazatec people of Oaxaca, S. divinorum has long been used in ceremonial healing rituals as a means of inducing a visionary state that allows the participants to divine the cause of illness or ailment (Wasson, 1962; Valdés et al., 1983). The psychotropic activity of this mint species has been attributed to a neoclerodane diterpene found in the leaves, salvinorin A (1) (Ortega et al., 1982; Valdés et al., 1984, 1987; Siebert, 1994). It has been demonstrated that salvinorin A is an extraordinarily potent drug in humans, exhibiting threshold effects at doses in the range of 200‑500 micrograms, making it the most potent natural hallucinogen ever studied (Siebert, 1994). The site of action and mechanism of pharmacological activity of 1 remain unknown. When submitted for major neurotransmitter and second messenger receptor site screening, 1 showed no significant binding to these sites at concentrations of 10‑5 M (Siebert, 1994). It has been suggested that the extraordinary potency of salvinorin A, and the absence of any known mechanism for its effects, could indicate that a new receptor system is involved in the activity of the compound (McKenna, 1996).

phytochemicalmol.gif - 3843 Bytes



        Salvinorin A has thus far been attributed uniquely to S. divinorum; no other natural source for this compound has been identified nor has it been synthetically produced. The current distribution of S. divinorum suggests that all existing stands of the plant have been intentionally cultivated by humans; no clearly wild populations of the species have been identified (Reisfeld, 1993). Furthermore, while the plant is readily propagated through asexual reproduction by rooting stem cuttings, it appears that it rarely if ever reproduces from seed in nature. It has been proposed that S. divinorum may in fact be a hybrid, resulting in substantially reduced fertility within the species (Reisfeld, 1993). The parentage for such a hybrid may remain unknown as no obvious candidates have been proposed as likely progenitors.


        Since its initial isolation, no analytical assay has been published in order to quantify precisely the amount of salvinorin A present in samples of S. divinorum or any other plant species. The purpose of the current study was to develop a sensitive quantitative assay to determine the amount of 1 present in leaf samples of S. divinorum. In addition, the method has been used to screen for this compound in other species.




General experimental procedures. Salvinorin A (1) used for standard curve development was purified from S. divinorum leaf extracts and authenticated by nuclear magnetic resonance spectroscopy and by thin layer chromatographic (TLC) comparison with authentic 1 (Valdés et al., 1994). All solvents used for extraction and chromatography were of high performance liquid chromatographic (HPLC) grade from Fisher Scientific (Fair Lawn, NJ, USA). Water used in HPLC mobile phase mixtures was distilled and subsequently filtered through a 0.22 ~im membrane (Millipore Corp., Bedford, MA, USA).


Plant materials. Asexually propagated plants of S. divinorum used in establishing the HPLC methods were purchased from Logee's Greenhouses (Danielson, CT, USA) and multiplied by rooting additional cuttings in the S.B. Penick Experimental Greenhouse at the Philadelphia College of Pharmacy and Science. Voucher specimens were deposited in the herbarium of the Philadelphia College of Pharmacy and Science. Additional S. divinorum samples used for analysis were collected from established outdoor stands.


Extraction of samples. Plant tissues were lyophilized after harvest and ground to a homogeneous powder in a Wiley mill (no. 20 mesh). Samples (0.500 g) of lyophilized whole leaf were extracted in 125 mL roundbottom flasks by steeping in 25 mL of chloroform for 30 min. The extract was filtered and the filtrate set aside. The extraction flask and filtered solids were rinsed with an additional 15 mL of fresh chloroform. The filtrate from the rinse was then combined with the original filtrate and the resulting solution was evaporated to dryness with a rotary evaporator. The dry solids were redissolved in a mixture of 20.0 mL methanol and 5.0 mL acetone using sonication to assist in dissolving of all solid material.


Chromatography conditions. HPLC was performed using a Milton Roy HPLC system (Riviera, FL, USA), consisting of a Constametric 3000 Series isocratic pump, a Rheodyne injector (Rheodyne L.P., Cotati, CA, USA), and a Spectromonitor 3100 variable wavelength UV‑VIS detector. The analogue detector output was acquired by an advanced computer interface (Dionex Corp., Sunnyvale, CA, USA), converted to a digital signal, and then processed by Al‑450 Chromatography Automation Software (Dionex). The HPLC column was a Zorbax (MACMOD Analytical, Inc., Chadds Ford, PA, USA) 300 SBC18 column (250 x 4.6 mm i.d.; 5 micrometer particle diameter; 300 A average pore size). In order to protect the integrity of the analytical column, all analyses were performed with a coupled C‑18 guard column. A mobile phase of acetonitrile:water (45:55) was used for all analyses at a flow‑rate of 1.0 mL/min, and 1 was detected by UV absorption at 208 mn.


Composite standard curve. A standard curve for 1 was produced using a solution of 0.051 mg/mL of standard dissolved in the HPLC mobile phase. By varying injection size, six different amounts of salvinorin A (0.255, 0.510, 0.765, 1.02, 1.28 and 1.53 [tg) were chromatographed: for each amount analyzed, three injections were made.


In order to validate the HPLC method, three separate determinations of the standard curve were performed. The data collected for each amount for all three curves (i.e. nine data points per amount) were averaged and replotted to yield a composite standard curve for salvinorin A.


Quantification of salvinorin A in plant tissue samples. For each sample, 20 microliter of the reconstituted methanol: acetone extract was injected into the HPLC and the area of the peak due to 1 was integrated. This peak area was used to calculate the amount of salvinorin A present in the tissue sample by applying the linear equation obtained from the composite standard curve.




Under the isocratic chromatographic conditions, authentic salvinorin A (1) eluted in approximately 8.0‑8.1 min (Fig. 1A). The analysis of the plant tissue extracts described below allowed rapid elution of polar components, baseline resolution of the analyte of interest, and a short analysis time (for examples, see Fig. 1).

phytochemicalfig1.gif - 13060 Bytes

Figure 1. HPLC chromatograms of an authentic sample (0.2551Ag) of salvinorin A (A), and of representative Salvia divinorum tissue extracts obtained from "Palatable" clone (Bret Blosser) (B), from Cerro Rab6n clone (L. J. Valdés) (C), and from a seed grown plant DS03 (D. J. Siebert) (D). In each case the retention time of the peak representing salvinorin A is indicated. (For chromatographic protocol see Experimental section).



phytochemicalfig2.gif - 6701 Bytes

Figure 2. The composite standard calibration for the quantification of salvinorin A by HPLC (the error bars indicate ±1 standard deviation; the linear regression equation for the calibration curve is y= 759,334x ‑44,127).




A composite standard curve for 1 was compiled from three individual determinations of the standard curve (Fig. 2). The correlation coefficients for these three standard curves were 0.9997, 0.9997 and 0.9967. The coefficient of variation among the three curves was 5.9%. The coefficients of variation for each injected amount across all three days ranged from 1.7 to 5.7%.


The peak identified as salvinorin A (1) from authentic standard solutions was observed in extracts of leaves and stem from S. divinorum. In order to examine the concentration range of 1 found in various populations of S. divinorum, 20 different samples of leaves were collected from cultivated plants in private collections as well as endemic populations of the plant in Oaxaca. Representative chromatograms for three different samples of S. divinorum are shown in Fig. 1 Estimations were based on the average of two separate injections. The determinations of the leaf content of salvinorin A showed a broad range of values ranging from 0.89 to 3.70 mg/g dry weight (Table 1). Stem material showed considerably lower values of 1 equivalent to approximately 4% of the level found in leaves.

phytochemicaltable1.gif - 8152 Bytes


Since the genetic heterogeneity of different populations of S. divinorum is unknown, it is not possible to attribute the different levels of 1 in these plants to any specific environmental or genetic conditions. Four separate collections of the plant which may represent distinct clones showed that variability in levels of 1 between clones is as great as the variability between plants of the same clone grown in different locations. These results are in contrast to earlier TLC analyses that suggested that there were no appreciable differences in levels of salvinorin A among plants grown in Michigan, Louisiana, or Mexico (Valdés, 1994). The Wasson and Hofmann clone samples showed a range of 1.94 to 3.70 mg/g of 1 with an average value of 2.69 mg/g (±0.67, n = 5). The "palatable" clone showed a range of 0.89 to 2.83 mg/g of 1 with an average of 1.73 mg/g (±0.99, n = 4). Two other samples from distinct collections made in Oaxaca contained 2.61 and 1.89 mg/g of salvinorin A.


An early report by Epling and Játiva (1962) showing that S. divinorum was most closely allied to S. concolor Lamb ex. Benth led us to screen this species for salvinorin A content. Ethnobotanical reports suggesting that Coleus blumei was similarly used as a psychotropic by the Mazatec (Wasson, 1962) caused us to screen one sample of this species as well for possible content of 1. HPLC analysis of the leaf extracts of these two species did not show any salvinorin A present, nor was 1 found in leaf extracts of S. blepharophylla, S. chiapensis, S. gregii var. San Isidro, S. leucantha, S. membranacea or S. recurva.


In summary, application of a novel HPLC methodology to quantify salvinorin A in S. divinorum leaves established the typical levels of this compound in a number of samples of the plant. Further screening of additional sage species for 1 may help elucidate the origins and taxonomic position of S. divinorum within the genus Salvia. Additionally, this analytical assay can be utilized in screening for salvinorin A in other S. divinorum plant parts.




We thank Dr. John R. Porter for advice on plant tissue culture and phytochemical methods and for critical reading of the manuscript, Dr. Anil D'Mello for his advice concerning the development of quantitative HPLC methodology, and Jeremy Mihalov, Anne Giordano and Mila Denisova for their invaluable contributions. We gratefully acknowledge PCPS for financial support, namely the Evelyn C. F. Ma Memorial Research Fellowship (J.W.G.) and Summer Research Salary Support (R.S.H.).



Epling, C. and Játiva‑M, C. (1962). A New Species of Salvia from Mexico. Bot. Mus, Leaflets, Harvard Univ. 20, 75‑76.


McKenna, D. J. (1996). Plant Hallucinogens: Springboards for Psychotherapeutic Drug Discovery. Behav. Brain Res. 73, 109‑115.


Ortega, A., Blount, J. F. and Manchand, P. S. (1982). Salvinorin, a New Trans‑Neocleroclane Diterpene from Salvia divinorum (Labiatae). J. Chem. Soc. Perkins Trans. 1,2505‑2508.


Reisfeld, A. S. (1993). The Botany of Salvia divinorum (Labiatae). SIDA 15, 349‑366.


Siebert, D. J. (1994). Salvia divinorum and Salvinorin A: New Pharmacologic Findings. J. Ethnopharmacology 43, 53‑56.


Valdés III, L. J. (1994). Salvia divinorum and the Unique Diterpene Hallucinogen, Salvinorin (Divinorin) A. J. Psychoact. Drugs 26, 277‑283.


Valdés III, L. J., Diaz, J. L. and Paul, A. G. (1983). Ethnopharmacology of Ska Maria Pastora (Salvia divinorum, Epling & Jitiva‑M)., J. Ethnopharmacology 7, 287312.


Valdés III, L. J., Butler, W. M., Hatfield, G. M., Paul, A. G. and Koreeda, M. (1984). Divinorin A, a Psychotropic Terpenoid, and Divinorin B from the Hallucinogenic Mexican Mint Salvia divinorum. J. Org. Chem. 49, 4716‑4720.


Valdés III, L. J., Hatfield, G. M., Koreeda, M. and Paul, A. G. (1987). Studies of Salvia divinorurn (Lamiaceae), a Hallucinogenic Mint from the Sierra Mazateca in Oaxaca, Central Mexico. Econ. Bot. 41, 283‑291.


Wasson, R. G. (1962). A New Mexican Psychotropic Drug from the Mint Family. Bot. Mus. Leaflets, Harvard Univ. 20,7784.


Parts of this work were presented at the 1996 American Chemical Society Mid‑Atlantic Regional Meeting, Student Award Symposium, The Chromatography Forum of Delaware Valley, May 24, 1996, Villanova University, Villanova, PA, Abstract no. 363.


Contract/grant sponsor: PCPS, the Evelyn C. F. Ma Memorial Research Fellowship.


Contrat/grant sponsor: PCPS, Summer Research Salary.