J. Sep. Sci. 2008, 31, 3959 –3964 C. S. Rumalla et al. 3959

Chidananda Swamy Rumalla1

Bharathi Avula1

Yatin J. Shukla2

Yan-Hong Wang1

Rahul S. Pawar1

Troy J. Smillie1

Ikhlas A. Khan1, 2

1National Center for NaturalProducts Research, ResearchInstitute of PharmaceuticalSciences, The University ofMississippi, University, MS38677, USA

2Department of Pharmacognosy,School of Pharmacy, TheUniversity of Mississippi,University, MS 38677, USA

Original Paper

Chemical fingerprint of Hoodia species, dietarysupplements, and related genera by using HPTLC

A HPTLC method was developed for simple and rapid chemical fingerprint analysisof four Hoodia species, dietary supplements that claim to contain Hoodia gordonii,and plants from genera related to Hoodia. HPTLC was performed on precoated silica60F254 plates with dichloromethane/methanol/water 75:17:2.2 by volume, as mobilephase. Evaluation of the HPTLC plates was done by using the CAMAG DigiStore2 dig-ital system with winCATS software. The authentication of H. gordonii was achievedby comparing the band colors and Rf values for TLC fingerprints with those of 11standard compounds including P57. The developed method was successfully appliedfor the identification of the 11 pregnane glycosides for four different species ofHoodia, 24 related genera and 13 dietary supplements that claim to contain H. gordo-nii. Different sample matrices were successfully analyzed, providing a wide range ofapplicability for this method, including gels, capsules, tablets, sprays, teas, snackbars, powders, and juices. The developed method was validated for specificity, stabil-ity, repeatability, and robustness. The results of HPTLC method were verified by LC-UV-MS method.

Keywords: Dietary supplements / Hoodia gordonii / Hoodia species / HPTLC / Pregnane glycosides /

Received: August 1, 2008; revised: September 22, 2008; accepted: September 22, 2008

DOI 10.1002/jssc.200800441

1 Introduction

Traditionally, Hoodia gordonii is used as an appetite sup-pressant by the San people of the Kalahari Desert [1]. TheSan hunters use fresh aerial parts of H. gordonii to sup-press hunger on their hunting trips [2, 3]. In 1998, theCouncil for Scientific and Industrial Research (CSIR) inSouth Africa patented the oxypregnane glycoside P57 [4],which was considered to be the active constituent of H.gordonii [5, 6]. In the last decade, the ethnobotanicalknowledge regarding the anorectic properties of thisplant has led to the formulation of several herbal weight-loss products claiming to contain H. gordonii. In the USmarket, Hoodia products are sold as OTC dietary supple-ments, and have emerged as one of the most sought-afterbotanicals for weight loss [7]. As H. gordonii is a slow-grow-ing succulent, the increased demand for raw plant mate-rial and products cannot be met by the reduced supply ofthis endangered plant. Consequently, there are moreconcerns about the possibility of adulteration, safety,and quality of Hoodia containing products. Therefore,developing a simple yet discerning method for the iden-

tification, and quality analysis of H. gordonii plant materi-als as well as products that claim to contain H. gordonii,has become crucial.

Previously, LC-ESI-TOF and LC-UV methods for thedetermination of P57 in Hoodia species plant materialand dietary supplements were established in our labora-tory [8]. A recent HPTLC method by Widmer et al. [9] hasalso been reported, which uses P57 as the marker for theidentification of Hoodia raw materials. Further, Widmeret al. [9] also suggested the use of inexpensive and com-monly found compounds like b-sitosterol, saccharose,and fructose as substitute reference standards. However,phytochemical studies have shown that H. gordonii pos-sesses a wide range of unique pregnane glycosides, com-prising of at least two chemically distinct steroidal agly-cones [10, 11]. Therefore, use of only P57, or use of non-specific compounds like b-sitosterol, saccharose, andfructose as markers, may not establish the completechromatographic identity of H. gordonii.

A recent study by our group [12] has reported HPLC-based chemical fingerprint analysis of H. gordonii, andrelated species, using 11 pregnane glycosides as markerconstituents. However, the need for sophisticated instru-ments and long run times restrict the applicability ofthis method for routine basis. These issues expressed theneed for a simpler, faster, and robust method.

Correspondence: Professor Ikhlas A. Khan, National Center forNatural Products Research, University of Mississippi, University,MS 38677, USAE-mail: ikhan@olemiss.eduFax: +1-662-915-7989

i 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com

3960 C. S. Rumalla et al. J. Sep. Sci. 2008, 31, 3959 – 3964

i 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com

Figure 1. Structures of steroidal glycosides.

J. Sep. Sci. 2008, 31, 3959 –3964 Other Techniques 3961

In our previous work [10, 11], pregnane glycosideswere isolated which were chemically related to P57:hoodigoside M (1), hoodigoside L (2), hoodigoside P (3),hoodigoside U (4), hoodigoside O (5), hoodigoside E (6),hoodigoside F (7), hoodigoside J (8), hoodigoside N (9),P57 (10), and hoodigoside C (11); structures are shown inFig. 1. In the present work, a fast and reliable HPTLCmethod was developed for the chemical fingerprint anal-ysis of 11 steroidal glycosides. The method is also usedfor the comparison studies of various related genera anddietary supplements. Different sample matrices weresuccessfully analyzed, providing a wide range of applic-ability of this method, including gels, capsules, tablets,sprays, tea bags, snack bars, powders, and juices.

2 Experimental

2.1 Chromatographic equipment

The HPTLC system (CAMAG, Muttenz, Switzerland) con-sisted of (i) DigiStore2 digital system along with winCATSsoftware ver. 1.4.3. (CAMAG); (ii) Linomat 5 sample appli-cator using 100 lL syringes, and connected to a nitrogentank. Twin trough chamber 20 cm610 cm (Analtech,USA), Sonicator (Fisher Scientific, USA), Millex 0.20 lmsyringe Driven filter Unit (Millipore, USA).

2.2 Chemicals and plant materials

The standard compounds (1–11) were isolated atNational Center for Natural Products Research (NCNPR)(Fig. 1). The identity and purity were confirmed by chro-matographic (TLC, HPLC) methods and the analysis of thespectroscopic data (IR, 1D- and 2D-NMR, HR-ESI-MS) [10,11].

Different Hoodia species and various related plant sam-ples: H. gordonii (voucher no. 2821), H. ruschii (voucher no.2822), H. currorii (voucher no. 2823), Asclepias labriformis(voucher no. 2614), Huernia keniensis (voucher no. 2818),Tridentea choanantha (voucher no. 2813), Stapelia flavirostris(voucher no. 2812), Edithocolea grandis (voucher no. 2815),Cynanchum marnieranum (voucher no. 2825), Heurnia recon-dita (voucher no. 2819), S. gigantea (voucher no. 2809), S.schinsii (voucher no. 2810), S. leendertziae (voucher no.2811), Piaranthus globosus (voucher no. 2814), C. perrieri(voucher no. 2824), Ceropegia dichtoma (voucher no. 2820),Orbea variegata (voucher no. 2817) were obtained fromMissouri Botanical Garden. H. parviflora (voucher no.3163) was obtained from AHP. Gonolobus condurango(voucher no. 1738), Marsdenia cundurango (voucher no.1598), a perported Hoodia spp. (voucher no. 2881), Vince-toxicum hirundinaria (voucher no. 42), and Caralluma fim-briata (voucher no. 2971) were obtained from commercialsuppliers. Hemidesmus indicus (voucher no. 1224) and Gym-nema sylvestris (voucher no. 1223) were obtained from SAI

Laboratories, Hyderabad, India. Opuntia ficus-indica (sam-ple no. 2888) was purchased from a local store. C. stratum(voucher no. 838), C. stauntonii (voucher no. 780) wereobtained from Beijing Yuke Botanical Development(China). Voucher specimens of all samples are depositedat the NCNPR, University of Mississippi, Mississippi, USA.

2.3 Sample preparation

Dry plant samples (0.3 g) or an adequate amount of pow-dered products including tablet/snack bars or capsule/tea bags/powder pack contents were weighed (averageweight of dosage forms) and sonicated in 2.5 mL of meth-anol for 20 min followed by centrifugation for 10 min at3300 rpm. The supernatant was transferred to a 10 mLvolumetric flask. The procedure was repeated threetimes and the respective supernatants were combined.The final volume was adjusted to 10 mL with methanoland mixed thoroughly. Prior to application, an adequatevolume (ca. 2 mL) was passed through a 0.2 lm nylonmembrane filter.

2.3.1 Juices/sprays

One milliliter of juice/spray solution was mixed with1.0 mL of methanol, vortexed for 30 s and sonicated for30 min. It was then vortexed and centrifuged for 15 minat 3300 rpm. The clear supernatant solution was used foranalysis.

2.4 Preparation of standard solutions

An individual stock solution of the standard compoundswas prepared at a concentration of 1.0 mg/mL in metha-nol and further diluted to 100 lg/mL. Fifty microliter ofeach standard solution was mixed together, and the mix-ture was used for HPTLC analysis (application volume22 lL).

2.5 HPTLC plates

Glass plates (Merck, Darmstadt, Germany) with silica gel60F254 (20 cm610 cm; 10 cm610 cm). Before use plateswere prewashed with methanol and dried in an oven for3 min at 1008C.

2.6 Sample application

Sample application with CAMAG Linomat 5, applicationposition from the lower edge of the HPTLC plate is10 mm at dosing rate 90 nL/s. Each plate accommodated(i) 15 tracks of sample of H. gordonii, standard mixture of11 marker compounds and individual marker standards(1–11), (ii) six tracks of samples of Hoodia species (H. gordo-nii, H. ruschii, H. currorii, and H. parviflora), (iii) 29 tracks ofsamples of H. gordonii and related genera, (iv) 16 tracks of

i 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com

3962 C. S. Rumalla et al. J. Sep. Sci. 2008, 31, 3959 – 3964

sample of H. gordonii and different dosage forms of diet-ary supplements. All samples were applied according tothe following settings: 8 mm from the bottom of theplate, band width 8 mm; distance between bands10 mm; application volume 1–4 lL. All remaining meas-urement parameters were left at default settings.

2.7 Development

Chamber saturation was done using 20 cm610 cmWhatman filter paper for 20 min. Development was indichloromethane/methanol/water (75:17:2.2 by volume),until 70 mm from the lower edge of the plate, tempera-ture 20–258C and relative humidity 55–65% which weremeasured by Acurite instrument.

2.8 Derivatization reagent

Anisaldehyde reagent was prepared by mixing 85 mL ofmethanol, 10 mL of acetic acid, and 5 mL of concentratedsulfuric acid. To this solution, 0.5 mL of p-anisaldehydereagent was added. Developed plates were immersed for2 s, and dried in an oven for 5 min at 1008C.

2.9 Documentation

CAMAG DigiStore2 digital system with winCATS soft-ware 1.4.3 was used for the documentation of derivatizedplates under white light.

2.10 Stability

Sample solutions of the extracts were prepared andstored at room temperature for 3 days and then appliedon the same HPTLC plates; after development the chro-matogram was evaluated for additional bands. Similarlybands stability was analyzed by keeping the resolvedpeaks and inspecting at intervals of 12 and 24 h.

3 Results and discussion

3.1 HPTLC fingerprinting identification

The chromatographic conditions, in particular the devel-oping solvents (i.e., types of solvents and ratios), werecarefully optimized before the H. gordonii sample wasanalyzed. The method was developed for the standardcompounds, then it was applied to H. gordonii plant mate-rial and botanical formulations claiming to contain H.gordonii. The results observed by the derivatization of thedeveloped plates with anisaldehyde reagent (many of thebands were not UV active), and those observed under thewhite light, showed good separation for all the com-pounds in H. gordonii. Standard compounds (pregnane

glycosides, 1–11) were separated on the same plates forthe authentication of H. gordonii sample.

The samples were identified by matching the colors,and Rf values of the bands with those of the standard com-pounds. These compounds appeared in the chromato-gram at significantly different Rf values. For positiveidentification, a sample must exhibit bands with chro-matographic characteristics, including colors, and Rf val-ues (0.08 for 1, 0.18 for 2, 0.20 for 3, 0.25 for 4, 0.31 for 5,0.41 for 6, 0.42 for 7, 0.46 for 8, 0.53 for 9, 0.62 for 10, and0.68 for 11), similar to those of standard compounds(Fig. 2 and Table 1). A good resolution and well definedbands were obtained when the chamber was saturatedwith the mobile phase for 20 min at room temperature.

3.2 Analysis of Hoodia species, related genera,and dietary supplements

The presence of the 11 pregnane glycosides was identifi-able within all Hoodia spp. with the outlined method. Theutilization of these marker compounds for HPTLC canprovide for the identification of Hoodia spp. Figure 3shows the comparison of different Hoodia species (H. gor-donii, H. currorii, H. ruschii, and H. parviflora), which showedthe presence of all 11 standard compounds, except forcompound 9, which was present in very minute amountin H. parviflora. Also, H. parviflora displayed an, as yet,unidentified band X that was more prominent in H. parvi-flora as compared to other species of Hoodia.

Once the chemical fingerprinting method had beendeveloped through the use of authenticated plant mate-rial, its applicability for the distinctive identification ofH. gordonii over other plant materials had to be tested. Allspecies of Hoodia and other related plant materials usedin the study are listed in the “Section 2.2”. Utilizing thismethod, we selected a number of closely related generabelonging to same family, Asclepiadaceae, and O. ficus-ind-ica which is a member of the Cactaceae family. Thesesamples can be easily confused with authenticated H. gor-

i 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com

Table 1. Standard compounds from H. gordonii and their Rf

value as determined by the HPTLC method

No. Compound name Rf

1 Hoodigoside M 0.082 Hoodigoside L 0.183 Hoodigoside P 0.204 Hoodigoside U 0.255 Hoodigoside O 0.316 Hoodigoside E 0.417 Hoodigoside F 0.428 Hoodigoside J 0.469 Hoodigoside N 0.53

10 P57 0.6211 Hoodigoside C 0.68

J. Sep. Sci. 2008, 31, 3959 –3964 Other Techniques 3963

donii, therefore the outlined method needs to be able toprovide a clear fingerprinting profile in order to differen-tiate these samples. Figure 4 details the fingerprint com-parison of Hoodia species with 24 plant genera, and spe-cies. All genera, when compared to H. gordonii, exhibitedvery different fingerprints. The sample in track number27 is a commercial product that claims to be Hoodia spp.,but its fingerprint does not match with that of authenti-cated H. gordonii or any of the 11 marker compounds,thus indicating that this sample is most likely not H. gor-donii.

Along with the authentication of plant materials,application of the present method was extended towardthe identification, and quality analysis of dietary supple-ments containing H. gordonii. Thirteen commerciallyavailable dietary supplements were tested (DS-1–DS-13).Figure 5 shows the comparison of H. gordonii with variousdietary supplements that was comprised of a wide rangeof formulation matrices such as gels, capsules, tablets,sprays, tea bags, snack bars, powders, and juices. Two(DS-1 and DS-2; tracks 4 and 5) products contained P57,and showed similar profiles as H. gordonii, which con-firmed the presence of H. gordonii in these dietary supple-ments. Eleven products (DS-3–DS-13) did not show thepresence of any of the 11 pregnane glycosides. Thesetracks (6–16) showed different chemical profiles, as com-pared to Hoodia species which may be attributed to thepresence of other botanicals in the dietary supplements.

By using this method we have selected a wide varietyof samples including those containing a complex mix-ture of Hoodia with other botanicals (Fig. 5; tracks 5, 7, 9–16) as well as Hoodia only samples (Fig. 5; tracks 4, 6, 8).Tracks 4–5 showed for the presence of Hoodia. If the sam-ple contains Hoodia then it will provide at least a partialmatch with the fingerprint profile of the standards mix-ture. Moreover the samples were also analyzed by UPLC-UV-MS [13] to confirm these results. LC-MS coupled withESI interface method was applied for the identificationof P57 in various Hoodia samples, related genera, and

i 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com

Figure 2. HPTLC chromatogram of H.gordonii and its standard compounds.Evaluation by HPTLC on silica gel 60 F254

with dichloromethane/methanol/water75:17:2.2 by volume as a mobile phase;derivatization with anisaldehyde reagent.Tracks: 1, 15, H. gordonii; 2,14, Std Mix1–11; 3, hoodigoside M; 4, hoodigoside L;5, hoodigoside P; 6, hoodigoside U; 7,hoodigoside O; 8, hoodigoside E; 9, hood-igoside F; 10, hoodigoside J; 11, hoodigo-side N; 12, P57; 13, hoodigoside C.

Figure 3. Comparison of different Hoodia species. Tracks: 1,H. gordonii; 2, H. currorii; 3, Std Mix-11; 4, P57; 5, H. ruschii;6, H. parviflora.

Figure 4. Comparison of H. gordonii with other related genera. Tracks: 1, 29, H. gordonii; 2, C. fimbriata; 3, O. ficus-indicus; 4,Orbea variegate; 5, C. stratum; 6, G. cundurango; 7, Std Mix-11; 8, 22, P57; 9, G. sylvestris; 10, M. cundurango; 11, A. labrifor-mis; 12, H. keniensis; 13, T. choanantha; 14, C. dichtoma; 15, S. flavirostris; 16, E. grandis; 17, C. stauntonii; 18, C. marniera-num; 19, V. hirundinaria; 20, H. indicus; 21, H. recondita; 23, S. gigantea; 24, S. schinsii; 25, S. leendertziae; 26, P. globosus;27, Hoodia spp.; 28, C. perrieri.

3964 C. S. Rumalla et al. J. Sep. Sci. 2008, 31, 3959 – 3964

dietary supplements [13]. This method involved the useof the [M + Na]+ ions in the positive ion mode withselected ion recording (SIR). In the positive ion mode,[M + Na]+ at m/z 901.5 for compound P57 was observed.The plant samples (H. gordonii, H. currorii, H. ruschii, and H.parviflora) and commercial products (DS-1 –DS-2) wereanalyzed and showed the presence of P57. Products (DS-3–DS-13), Opuntia sample and non-Hoodia plant samplesdid not show the presence of P57 using this method.Hence, these results show that the current HPTLCmethod can be successfully used to qualitatively analyzea broad range of formulations, for the presence orabsence of H. gordonii.

The validation parameters for the developed methodwere specificity, stability, precision (repeatability), androbustness. Overlapping of bands pose a typical chal-lenge for HPTLC in the case of complex samples such asHoodia. In such a scenario, the use of multiple markercompounds helps to overcome this problem, andincreases the specificity of the method. The presentmethod was found to be specific, as there was no interfer-ence of substances in the regions of interest, i.e., thebands of standard compounds. Sample solutions of theextract were found to be stable on the HPTLC plates for atleast 3 days. No additional bands were observed afterderivatization of the HPTLC plates and the chromato-gram was found to be stable. Repeatability was deter-mined by running a minimum of five analyses for themixture of standard compounds. The Rf values for thestandard compounds were calculated with repeatability(n = 5). Rf values for 11 standard compounds varied lessthan 0.02 (% RSD was in the range of 1.05–5.73). By intro-ducing small changes in the mobile phase composition,mobile phase volume, duration of mobile phase satura-tion, and activation of prewashed HPTLC plates withmethanol, the effects of the results were minute. The

robustness of the method was critically dependent onthe type of plate used (TLC or HPTLC) and was sensitive tochanges in temperature and relative humidity.

4 Concluding remarks

The HPTLC fingerprinting method has the advantages ofsimplicity, rapidity, and visuality. The method is suitablefor rapid, decisive authentication, and comparison of thesubtle differences among samples from various plantsources. The chemical fingerprint method developed inthis study was able to verify the presence of Hoodia spe-cies in raw materials, various matrices or if the material/product contains other genus. The developed HPTLCmethod was validated for specificity, reproducibility,and robustness. It provides a faster, cost effective alterna-tive for qualitative analysis, and thus can be useful in theauthentication of H. gordonii. LC-UV-MS confirmation wasdone for all the samples analyzed by HPTLC.

This research is funded in part by “Science-Based Authenticationof Dietary Supplements” funded by the Food and Drug Adminis-tration grant number 2 U01 FD 002071-07. The authors wouldlike to thank the Missouri Botanical Garden, Missouri, USA andAHP, USA and Beijing Yuke Botanical Development Co. Ltd.,China for providing plant samples.

The authors declared no conflict of interest.

5 References

[1] Van Heerden, F. R., Horak, M. R., Maharaj, V. J., Vleggaar, R., Sen-abe, J. V., Gunning, P. J., Phytochemistry 2007, 68, 2545 – 2553.

[2] Anonymous, Consumer Rep. 2006, 71, 49.

[3] Anonymous, Bulletin of the World Health Organization 2006, 84, 345.

[4] Van Heerden, F. R., Vleggaar, R. M., Horak, R., Learmonth, R. A.,Maharaj, V., Whittal, R. D., International Patent, WO 98-GB1100,9846243, 162, CSIR, South Africa 1998.

[5] MacLean, D. B., Luo, L.-G., Brain Res. 2004, 1020, 1 – 11.

[6] Habeck, M., Drug Discov. Today 2002, 7, 280 – 281.

[7] Anonymous, Nutr. Bus. J. 2007, 9,12 – 14.

[8] Avula, B., Wang, Y. H., Pawar, R. S., Shukla, Y. J., Schaneberg, B.,Khan, I. A., J. AOAC Int. 2006, 89, 606 – 611.

[9] Widmer, V., Reich, E., DeBatt, A., J. Planar. Chromatogr. – Mod. TLC2008, 21, 21 – 26.

[10] Pawar, R. S., Shukla, Y. J., Khan, S. I., Avula, B., Khan, I. A., Steroids2007, 72, 524 – 534.

[11] Pawar, R. S., Shukla, Y. J., Khan, I. A., Steroids 2007, 72, 881 – 891.

[12] Avula, B., Wang, Y. H., Pawar, R. S., Shukla, Y. J., Khan, I. A., J.AOAC Int. 2007, 90, 1526 – 1531.

[13] Avula, B., Wang, Y. H., Pawar, R. S., Shukla, Y. J., Smillie, T. J.,Khan, I. A., J. Pharm. Biomed. Anal. 2008, 48, 722 – 731.

i 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com

Figure 5. Comparison of H. gordonii with dietary supple-ments of various dosage forms. Tracks: 1, H. gordonii; 2, StdMix-11; 3, P57; 4, 6, 12–16, capsules of DS-1, DS-3, DS-9–DS-13; 5, 7, tablets of DS-2 and DS-4; 8, juice of DS-5; 9,snack bar (DS-6); 10, tea bag of DS-7; 11, spray of DS-8.