Chemical characterisation of Hoodia gordonii extract

Food and Chemical Toxicology 50 (2012) S6–S13

Contents lists available at ScienceDirect

Food and Chemical Toxicology journal homepage: www.elsevier.com/locate/foodchemtox

Chemical characterisation of Hoodia gordonii extract P.J. Russell a,⇑, C. Swindells b a b

Safety & Environmental Assurance Centre, Unilever, Colworth Science Park, Sharnbrook, Bedford, MK44 1LQ, UK Phytopharm plc, Lakeview House, 2 Lakeview Court, Ermine Business Park, Huntingdon, Cambridgeshire PE29 6UA, UK

a r t i c l e

i n f o

Article history: Received 19 January 2010 Accepted 22 February 2011

Keywords: Hoodia gordonii Steroid glycosides Mass spectrometry HPLC SDS–PAGE

a b s t r a c t The chemical composition of a solvent extract of Hoodia gordonii termed ‘H. gordonii extract’ has been characterised by hyphenated chromatographic methods and traditional analytical techniques. The extract consists of a mixture of steroid glycosides, fatty acids, plant sterols and polar organic material. High performance liquid chromatography (HPLC) with ultra violet (UV) and mass spectrometric (MS) detection was used to quantify and confirm the identity of a number of steroid glycosides (73.7% w/w) present in the extract. Gas chromatography (GC) with MS and flame ionisation detection (FID) was applied to determine the fatty acid (3.12% w/w) sterol (0.39% w/w) and alcohol (0.03% w/w) content of a saponified sample of the extract. Polar organic material was quantified by gravimetric methodology using C18 SPE separation and was determined to be a minimum of 3% w/w. Moisture content was measured by Karl Fischer coulometric titration (0.81% w/w). The protein content was investigated by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) with SYPRO Ruby staining and a negative result was determined with a limit of detection of <0.001%w/w of protein per band. The chemical composition of the extract remained stable for 19 months when stored in re-sealable plastic bags at ambient (21–24 °C) temperature and <60% relative humidity. Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction Hoodia gordonii (Masson) Sweet ex Decne is a perennial, succulent plant species from the Apocycnacaea family (previously the Asclepiadaceae family) indigenous to the arid regions of South Africa, Botswana and Namibia (Bruyns, 2005). In the past, several species of Hoodia were included in a research program undertaken by the National Food Research Institute of the Council for Scientific and Industrial Research (CSIR, South Africa). Their research showed that a specific molecule present in H. gordonii (described here as H.g.-12) decreased food intake and body weight in animals (van Heerden et al., 2007) and subsequently extracts of the plant are being investigated for the dietary control of body weight in humans. To date there have been no further published studies to suggest activity from other individual components in H. gordonii.

Abbreviations: HPLC, high performance liquid chromatography; UV, ultra violet; MS, mass spectrometric; GC, gas chromatography; FID, flame ionisation detection; SDS–PAGE, sodium dodecyl sulphate–polyacrylamide gel electrophoresis; GLP, good laboratory practice; BSTFA, N,O-bis (trimethylsilyl) trifluoroacetamide; TMCS, trimethylchlorosilane; MSD, mass spectrometric detector; APCI, atmospheric pressure chemical ionisation; FAME, fatty acid methyl esters; PBS, phosphate buffered saline; PAH, polycyclic aromatic hydrocarbons. ⇑ Corresponding author. Tel.: +01234 264831; fax: +01234 264744. E-mail address: [email protected] (P.J. Russell). 0278-6915/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.fct.2011.02.020

Recent studies suggest that approximately 300,000 deaths per year in the USA are attributed to obesity (Cheetham et al., 2004). Economic and technological changes promoting a sedentary lifestyle with easy access to low-cost, high-calorific food are factors fuelling this epidemic (Harrold et al., 2004). A more balanced diet, associated with exercise and with decreased food intake assisted by an efficacious H. gordonii product will be beneficial for reducing and controlling body weight. H. gordonii extract was obtained by methanol extraction of dried whole H. gordonii plant (all aerial parts but excluding the roots), followed by purification steps with heptane and heptane/ methyl ethyl ketone (MEK) (Knight et al., 2012). Plants were botanically verified as H. gordonii, based on floral characteristics, by a CSIR asclepiad expert (Peckover, personal communication) and a voucher specimen was deposited in the herbarium of the South African National Biodiversity Institute (SANBI) in Pretoria. Comprehensive chemical characterisation of the resulting dried extract was necessary to determine the material composition for use in a variety of safety studies initiated as part of the development programme (Dent et al., 2012a,b). The key analytical results reported in this publication were obtained on a single typical batch of H. gordonii extract using validated methodologies within Good Laboratory Practice (GLP) accredited laboratories. It should be noted that the results reported here are from analysis of a specific H. gordonii extract from a known, verified source and whilst the

P.J. Russell, C. Swindells / Food and Chemical Toxicology 50 (2012) S6–S13

S7

3. Methods

peaks eluting within a specified window and has been successfully applied within this study. In this paper, seven individual steroid glycosides are considered in addition to the total. These steroid glycosides are referred to as H.g.-12, H.g.-21, H.g.-20, H.g.-23, H.g.-24, H.g.-17 and H.g.-19 in Janssen et al., in order of their elution in the HPLC chromatogram. H.g.-12 is equivalent to the steroid glycoside identified as Compound 1 (P57AS3) by Van Heerden et al. which was considered to be efficacious in a reduction of food intake in rats (van Heerden et al., 2007). The methodology was considered superior to that reported by Avula et al. (2006) in that it demonstrated improved resolution and hence quantitation for the later eluting steroid glycosides (i.e. H.g.-17 & H.g.-19). H. gordonii extract was accurately prepared at a concentration of ca 1.0 mg/ml (n = 9) in acetonitrile/water (1/1 v/v) to determine the concentration of steroid glycosides and assess sample homogeneity. Steroid glycoside concentrations were measured by High Performance Liquid Chromatography (HPLC) with Ultra Violet (UV) detection using an Agilent 1100 HPLC system (Agilent Technologies, Wokingham, UK). Steroid glycosides were defined as all peaks eluting after 15 min that were not present in the blank acetonitrile/water (1/1 v/v) sample. Common MS fragment ions at m/z 295 and 313 along with common glycoside losses were used to confirm that the minor peaks were structurally related to H.g.12. The HPLC system consisted of a solvent degasser, binary pump, autosampler, column compartment and multi-wavelength detector. Retention time comparisons were made with the Janssen et al. method and component identity was tentatively confirmed by single quadrupole LC–MS using an Agilent Technologies D Grade single quadrupole mass spectrometric detector (MSD), (Agilent Technologies, Wokingham, UK). Relative response factors based on molecular weight were used to quantify the steroid glycosides in the H. gordonii extract against a H.g.-12 calibration. H.g.-12 standards were prepared in duplicate using separate weighings from 5 to 100 lg/ml in acetonitrile/water (1/1 v/v). The solutions were analysed by injecting 20 ll onto a Zorbax RX-C8 (250  4.6 mm, 5 lm) analytical column with a Zorbax RX-C8 (12.5  4.6 mm, 5 lm, both Agilent Technologies) guard column held at 40 °C. A gradient flowing at 1 ml/min was used for the analysis with mobile phase starting conditions of 41% acetonitrile/ methanol (85/15 v/v) and 59% water/methanol (85/15 v/v). Initial conditions were held for 10 min before being linearly increased to 88% acetonitrile/methanol (85/15 v/v) and 12% water/methanol (85/15 v/v) over 30 min. The mobile phase was held at this composition for 5 min and then equilibrated back to the starting conditions. Steroid glycoside concentrations were quantified by LC–UV at 220 nm. The mass spectrometer was operated in positive ion atmospheric pressure chemical ionisation (APCI) mode, with ion source gas temperature 300 °C, gas flow 8 l/min (nitrogen), vaporiser temperature 400 °C, capillary voltage 4000 V and fragmentor 20 V. Scan range was m/z 90–1500.

3.1. Quantification and identification of steroid glycosides by LC–UV– MS

3.2. Quantification by GC-FID and GC–MS profile after saponification, methylation and silylation

A previously published paper (Janssen et al., 2008) reported the HPLC–UV/MS quantitation and identification of steroid glycosides in H. gordonii extracts, using a steroid glycoside external standard and relative response factors to relate other steroid glycosides to this standard. Relative response factors were assessed through a ring-trail of five independent laboratories which were actively applying this method using different HPLC equipment and they were found to be sufficiently robust for universal application (Unilever R&D Vlaardingen (Internal Report), 2008). This approach allowed the quantitation of individual peaks, and also the determination of a ‘‘total steroid glycosides’’ figure, covering all

H. gordonii extract was saponified with sodium hydroxide solution in methanol followed by methylation with boron trifluoride– methanol complex (20% solution). The method was designed to measure fatty acids from all sources, e.g. triglycerides, sterol esters and wax esters and was also used to quantify any sterols, alcohols and alkanes present. All compounds present were quantified against an undecanoic acid internal standard.

techniques described could undoubtedly be employed to assess potential adulteration, contamination or mis-identification of H. gordonii this was not the primary aim.

2. Chemicals and reagents H. gordonii extract was prepared from washed whole plants (aerial parts) which were comminuted to facilitate drying to a moisture level of less than 8%. Material was then extracted with methanol (50–80 °C) and plant waxes were removed through cooling, filtration, concentration and washing with n-heptane. The concentrated aqueous extract containing the steroid glycosides was subjected to four extractions with heptane and methyl–ethyl ketone (MEK), the organic phases collected and combined. The combined extract was treated with an aqueous wash of ethylenediamine tetra-acetic acid (EDTA) and sodium hydroxide followed by a water wash to remove contaminant metals and secondary plant components. The organic phase was treated with active carbon to remove any polycyclic aromatic hydrocarbon (PAH’s), the solvent removed in vacuo and the H. gordonii extract isolated and dried (Knight et al., 2012). Acetonitrile (HPLC grade) and hexane (glass distiled grade) was purchased from Rathburn Chemicals Ltd. (Walkerburn, UK). Methanol (Reidel de Haen LC–MS grade), pyridine (anhydrous 99.8%), Hydranal Working Medium K diluent and Hydranal 5 K titrant were all sourced from Sigma–Aldrich (Poole, UK). Water was obtained from an in-house Milli-Q water purification system. Sodium chloride (HPLC grade) was obtained from Fisher Scientific (Loughborough, UK). Sodium hydroxide and sodium sulphate (AnalaR grade) were obtained from VWR (Lutterworth, UK). Ethyl acetate (ACS grade), phosphate buffered saline (PBS, pH 7.4), sodium dodecly sulphate (SDS, electrophoresis grade, P98.5%), boron trifluoride–methanol complex (14% solution), undecanoic acid free fatty acid internal standard, C4–C24 fatty acid methyl ester (FAME) and n-paraffin standard mixes were all purchased from Sigma–Aldrich (Poole, UK). N,O-Bis (trimethylsilyl) trifluoroacetamide (BSTFA) containing 10% trimethylchlorosilane (TMCS) derivatising agent was obtained from Perbio (Cramlington, UK). Reagents used for SDS PAGE analysis were XT sample buffer, XT reducing agent, Criterion XT gel, Bis–Tris 12%, 1X XT MES and MOPS running buffers, Broad Range Molecular Mass Markers and SYPRO Ruby stain which were all purchased from BIO-RAD (Hemel Hempstead, UK). H.g.-12 (P92.4%) single steroid glycoside reference standard was supplied by Phytopharm plc (Huntingdon, UK). The structure was confirmed by MS and NMR spectroscopy to be consistent with that described by van Heerden et al. (2007).

3.2.1. Saponification, methylation and silylation Approximately 50 mg of H. gordonii extract was accurately weighed into a flask containing 1.2 mg of undecanoic acid free fatty

S8

P.J. Russell, C. Swindells / Food and Chemical Toxicology 50 (2012) S6–S13

acid internal standard. This was saponified with 8 ml of a 0.5 mol/L sodium hydroxide solution in methanol for 30 min under reflux, shaking occasionally. The saponified sample was methylated with 10 ml of boron trifluoride–methanol complex (14% solution) for 4 min under reflux. Hexane (6 ml) was added to extract the components of interest under reflux for 4 min. The flask was allowed to cool and filled to approximately 75% with hexane and shaken, before filling to the top of the round bottom flask with a saturated aqueous solution of sodium chloride to facilitate the removal and separation of the hexane layer. As much as possible of the hexane layer was removed and dried with 0.5 – 1 g of sodium sulphate before evaporating 2 ml under nitrogen to leave approximately 50 ll. At this point, 200 ll of BSTFA containing 10% TMCS and 800 ll of pyridine were added. The vials were capped and heated in an oven at 60 °C for 30 min. The derivatised extracts were diluted with 1 ml of hexane and transferred to vials for analysis by GC-FID and GC– MS as below. In addition to the H. gordonii extract, a hexane blank control was taken through this procedure. Samples were analysed against C4–C24 fatty acid methyl esters (FAME) and an n-paraffin standard to aid peak identification.

3.4. Moisture content by Karl Fischer titration The moisture content of H. gordonii extract was determined by Karl Fischer coulometric titration using a fully automated titrator (Metrohm Titrando, Buckingham, UK). The standard Karl Fischer protocol was followed (Scholz, 1984) using Hydranal Working Medium K diluent and Hydranal 5 K titrant. The H. gordonii extract (1 g) was introduced directly into the titration cell with no prior dissolution and results were quantified against a water standard and reference blank. 3.5. Stability The stability of the extract was assessed by quantitation of the steroid glycoside content (HPLC–UV/MS) and monitoring the moisture content by Karl Fischer coulometric titration. Stability analyses were conducted at approximately 6, 11, 16 and 19 months with the sample stored in re-sealable plastic bags at ambient (21–24 °C) temperature and <60% relative humidity. 3.6. Protein analysis

3.2.2. GC-FID The samples were quantified by GC-FID (HP6890, Hewlett Packard, UK). The carrier gas was helium at a constant flow of 1 ml/min. The samples were separated on a CPSil 8CB analytical column (30 m  0.25 mm i.d., 0.25 lm) from Varian (Yarnton, Oxford, UK) with a pre-column of deactivated fused silica (1 m  0.53 mm). A temperature gradient starting at 65 °C held for 2 min then increasing to 280 °C at a rate of 10 °C/min, with a second ramp at 4 °C/min to 360 °C and a final hold time of 25 min was used. The samples were analysed in on-column mode with a volume of 1 ll injected. The FID was held at 380 °C (H2 40 ml/min, air 450 ml/min and make-up N2 45 ml/min). 3.2.3. GC–MS The samples were also analysed by GC–MS for peak identification (6890, Agilent Technologies, UK). The carrier gas was helium at a constant flow of 1 ml/min. The samples were separated by a temperature gradient starting at 65 °C held for 2 min then increasing to 280 °C at a rate of 10 °C/min. It was then increased to 325 °C at 4 °C/min and held for a further 20 min before re-equilibrating at the initial temperature. Separation was carried out on a VF5-MS analytical column (30 m  0.25 mm i.d., 0.25 lm) from Varian (Yarnton, Oxford, UK). The samples were analysed in splitless injection mode at a temperature of 250 °C with a volume of 1 ll injected (split after 1 min at 100 ml/min). The MS was operated in electron impact ionisation mode over the range m/z 29–800. 3.3. Gravimetric quantification of polar compounds H. gordonii extract was dissolved in 50 ml water/acetonitrile/ methanol (50/35/15 v/v/v) and sonicated for approximately 2  15 min before being filtered through a pre-weighed filter paper. The resulting filtrate was loaded onto a pre-conditioned Varian MEGA BE C18, 10 g, 60 ml SPE cartridge (Varian Ltd., Yarnton, Oxford, UK), and collected under vacuum. Subsequent 5 ml volumes of water/acetonitrile/methanol (50/35/15 v/v/v) were used to elute components from the SPE cartridge and collected. Aliquots (<0.5 ml) of each of these fractions were analysed using the LC– UV method reported above and compared to the unfractionated H. gordonii extract. The polar material eluting at the beginning of the chromatogram was defined as having a retention time cut-off point at 15 min. The fractions confirmed as containing early eluting material only were pooled, rotary evaporated and freeze dried before being weighed.

3.6.1. Extraction Approximately 2 g of H. gordonii extract was dissolved in 10 ml ethyl acetate. Volumes of 0.5 ml were taken and either [A] water or [B] 0.5 ml of ethyl acetate added prior to centrifugation (15 min, 10,500g). Samples [B] had the ethyl acetate supernatant removed and discarded and the residue re-suspended in water. The water phases from [A] and [B] extracts were dialysed against phosphate buffered saline (PBS) using Slide-A-Lyser dialysis cassettes (Perbio, Cramlington, UK). H. gordonii extract was also prepared as a solution in acetonitrile/water (50/50 v/v) and PBS + 2% SDS at a concentration of 91 mg/ml and mixed for 30 min followed by centrifugation at 20,000g for 5 min. 3.6.2. SDS–PAGE Criterion XT gel, Bis–Tris 12%, 12 + 2 well were placed in the mini-tank according to the manufacturer’s instructions. The tank was then filled to the mark with 1X XT MES (low molecular weight) or MOPS (high molecular weight) running buffer (20 diluted with water and chilled prior to use). The extracts prepared above were mixed in the proportions 5:5:1:9 acetonitrile/water (50/50 v/v) and PBS + 2%SDS extracts:4X sample buffer:20X reducing agent:water respectively or 14:5:1 (excluding water for the ethyl acetate extracts). The Broad Range Molecular Mass Markers were diluted 1:20 in reducing XT sample buffer. The H. gordonii extract samples and markers were placed in a boiling water bath for 5 min. The molecular mass markers and samples were then loaded (10 ll) into the wells of the gel using the manufacturers sample loading guide. The protein components were then electrophoretically separated using a BIO-RAD powerpack (Hemel Hempstead, UK) at 200 V constant with a limit of 15 W. All gels were run until the buffer front had eluted off the bottom of the gel. 3.6.3. Gel staining The gel was fixed and washed with 2  100 ml volumes of 10/ 90 MeOH/7% v/v acetic acid (aq) for 15 min each before being stained with 75 ml SYPRO Ruby luminescent stain. The gel was microwaved for 2  30 s on full power with a 30 s agitation period between. After a further 5 min agitation, the gel was microwaved on full power for a final 30 s. The gel was then agitated for a further 23 min and de-stained with 100 ml 10/90 MeOH/7% v/v acetic acid (aq) for 30 min and then rinsed with 100 ml water before being imaged for molecular weight analysis using a Typhoon 8600 Variable

S9

P.J. Russell, C. Swindells / Food and Chemical Toxicology 50 (2012) S6–S13

Mode Imager System, (GE HealthCare, Chalfont St Giles, UK). The gels were dried using the Gelair Drying System (BIO-RAD, Hemel Hempstead, UK) according to the manufacturer’s instructions.

RSD), in which a-tocopherol, although not a sterol, is also included. The C28H58O long-chain alcohol was identified (0.03% w/w) and it is likely that it is conjugated with long chain fatty acids to form wax esters.

3.7. Contaminant screening 4.1.2. Polar material As part of the characterisation process a suite of contaminant screens were performed to determine levels of potential contaminants present in H. gordonii extract. This included analysis for polycyclic aromatic hydrocarbons (PAHs), aflatoxins, trace metals, residual solvents (methanol, heptane, MEK) and pesticides. Although full details of the contaminant analyses are not reported here, the analytical methods employed were sufficiently validated and performed to UKAS and/or GLP standard. A list of the contaminants analysed is given in Table 2.

A minimum figure of 3.1% w/w (±6.6% RSD) was obtained for material eluting in <15 min in the HPLC–UV chromatogram. 4.1.3. Protein analysis

4. Results

SYPRO Ruby staining successfully visualised the molecular mass markers and no protein was found in any of the extractions prepared using either high or low molecular weight range gels. The detection limit for the SYPRO Ruby staining method was <0.001% w/w protein per band.

4.1. LC–UV/MS

4.1.4. Stability

A summary of quantitative results are shown in Table 1 and the sub-sample analysis for steroid glycoside content demonstrated extract homogeneity.

It was noted that H. gordonii extract stored sealed under ambient conditions ranging from 18–24 °C and up to 60% humidity did not show appreciable amounts of degradation over the course of 19 months when analysed by LC–UV at various intervals during this time period. Moisture levels also remained constant throughout this period.

4.1.1. Quantification by GC-FID and GC–MS profile after saponification, methylation and silylation The main compound classes in the FID chromatogram were identified by MS as fatty acids, sterols and alcohols. H. gordonii extract contained fatty acids as free acids or conjugated to form triglycerides or wax esters consisting mainly of C14:0 (myristic acid), C15:0 (pentadecylic acid), C16:0 (palmitic acid), C18:0 (stearic acid), C18:1 (oleic acid) and C18:2 (linoleic acid) with known and unknown fatty acids quantified as 2.76% and 0.36% w/w respectively giving a total content of 3.12% w/w (±7.21% RSD). The sterols have been identified as cholesterol, b-sitosterol, and stigmasterol and were quantified as a total of 0.39% w/w (±12.1%

5. Discussion The LC–UV chromatogram (Fig. 1) contained a large number (>30) peaks eluting in the 15–45 min region which gave a response at the non-specific wavelength of 220 nm. Therefore MS spectral interpretation was required for the tentative confirmation of identity of the steroid glycosides. Positive ion APCI-MS spectra show a dehydrated parent ion and characteristic fragmentation patterns, created by in-source fragmentation, down to the steroid backbone.

Table 1 Results of quantitative LC–UV analysis (220 nm) of Hoodia gordonii extract. Analyte Janssen et al. H.g.-12

Chemical name

3b-[b-D-thevetopyranosyl-(1 ? 4)-b-D-cymaropyranosyl-(1 ? 4)-b-Dcymaropyranosyloxy]-12b-tigloyloxy-14b-hydroxypregn-5-en-20-one H.g.-21 3-O-[b-D-oleandropyranosyl-(1 ? 4)-6-deoxy-3-O-methyl-b-Dallopyranosyl-(1 ? 4)-b-D-cymaropyranosyl-(1 ? 4)-b-D-cymaropyranosyl]12b-tigloyloxy-14b-hydroxypregn-5-en-20-one H.g.-20 3-O-[b-D-cymaropyranosyl-(1 ? 4)-6-thevetopyranosyl-(1 ? 4)-b-Dcymaropyranosyl-(1 ? 4)-b-D-cymaropyranosyl]- 12b-tigloyloxy-14bhydroxypregn-5-en-20-one H.g.-23 3-O-[b-D-oleandropyranosyl-(1 ? 4)-b-D-digitoxopyranosyl-(1 ? 4)-b-Dcymaropyranosyl-(1 ? 4)-b-D-cymaropyranosyl]- 12b-tigloyloxy-14bhydroxypregn-5-en-20-one H.g.-24 3-O-[b-D-oleandropyranosyl-(1 ? 4)-b-D-cymaropyranosyl-(1 ? 4)-b-Dcymaropyranosyl-(1 ? 4)-b-D-cymaropyranosyl]- 12b-tigloyloxy-14bhydroxypregn-5-en-20-one H.g.-17 3-O-[b-D-cymaropyranosyl-(1 ? 4)-b-D-cymaropyranosyl-(1 ? 4)-b-Dcymaropyranosyl-(1 ? 4)-b-D-cymaropyranosyl]- 12b-tigloyloxy-14bhydroxypregn-5-en-20-one H.g.-19 3-O-[b-D-oleandropyranosyl-(1 ? 4)-b-D-cymaropyranosyl-(1 ? 4)-b-Dcymaropyranosyl-(1 ? 4)-b-D-cymaropyranosyl-(1 ? 4)-b-Dcymaropyranosyl]- 12b-tigloyloxy-14b-hydroxypregn-5-en-20-one Aglycone 12–O-tigloyl-3b, 12b, 14b-trihydroxy-pregn-5-en-20-one Other steroid glycosides** Total steroid glycosides * **

Response Factor*

Relative retention time*

Mean steroid glycoside content (% w/w) (n = 9)

Mean steroid glycoside content after 19 months (% w/w) (n = 3)

1.000

1.00

10.2 ± 0.14

10.0 ± 0.05

1.164

1.09

3.59 ± 0.14

3.70 ± 0.01

1.164

1.12

6.27 ± 0.11

6.37 ± 0.03

1.130

1.19

6.21 ± 0.09

6.35 ± 0.02

1.146

1.29

7.81 ± 0.08

7.56 ± 0.04

1.146

1.33

5.16 ± 0.05

4.93 ± 0.02

1.309

1.40

2.55 ± 0.05

2.41 ± 0.02

0.490 1.081

0.70 –

1.46 ± 0.26 30.5 ± 1.38 73.7 ± 1.73

1.32 ± 0.01 29.9 ± 0.14 72.6 ± 0.18

With respect to steroid glycoside H.g.-12. Other steroid glycosides are defined as all additional peaks eluting after 15 min that are not present in the blank.

S10

P.J. Russell, C. Swindells / Food and Chemical Toxicology 50 (2012) S6–S13 Table 2 Results of contaminant screening of Hoodia gordonii extract. Component

Result

Residual solvents Methanol Heptane Methyl–ethyl ketone (MEK)

<140 ppm <500 ppm <260 ppm

Aflatoxins B1 B2 G1 G2 Total (B1 + B2 + G1 + G2)

<1.0 ppb <1.0 ppb <1.0 ppb <1.0 ppb <4.0 ppb

Pesticide residues Carbaryl Fenthion Dithiocarbamates (CS2)

<1.0 ppm <0.1 ppm 0.07 ppm

Polyaromatic hydrocarbons Naphthalene Acenaphthylene Acenaphthene Phenanthrene Anthracene Fluoranthene Pyrene Benzo(c)fluorene Cyclopenta(cd)pyrene# Benz(a)anthracene Chrysene 5-methyl chrysene Benzo(b)fluoranthene Benzo(j)fluoranthene+ Benzo(k)fluoranthene+ Benzo(a)pyrene Indeno(123,cd)pyrene Benzo(ghi)perylene Dibenz(ah)anthracene Dibenz(al)pyrene Dibenz(ae)pyrene Dibenz(ai)pyrene Dibenz(ah)pyrene Total

2.00 ppb 3.12 ppb 8.02 ppb 0.33 ppb <0.50 ppb <0.50 ppb <0.50 ppb <0.50 ppb 1345 ppb

Metals Mercury Copper Iron Chromium Nickel Zinc Cadmium Lead Arsenic

<0.1 ppm 60 ppm 54 ppm 11 ppm 7 ppm 8 ppm <0.1 ppm 0.1 ppm 0.7 ppm

1010 ppb 6.10 ppb 22.4 ppb 160 ppb 5.65 ppb 41.9 ppb 54.2 ppb 1.80 ppb 1.29 ppb 3.61 ppb 15.8 ppb 0.86 ppb 4.94 ppb 3.82 ppb

The MS fragmentation pattern for H.g.-12 is shown in Fig. 2 with a dehydrated parent ion [M H2O+H]+ of m/z 861. Subsequent neutral losses of 160 and 144 amu correspond to the thevetose and cymarose moieties respectively. Mass losses of 100 amu correspond to the tiglate detachment. The tiglated steroid backbone ion is observed at m/z 413 with the corresponding detiglated structure seen at m/z 313. Six other steroid glycosides were tentatively identified, in order of elution as H.g.-21, H.g.-20, H.g.-23, H.g.-24, H.g.-17 and H.g.-19, along with the aglycone. These molecules showed mass spectra molecular ion fragmentation consistent with the patterns seen for H.g.-12. The MS data used in conjunction with relative retention time and elution order confirmed that the identities of the key markers were consistent with the structures shown in Fig. 3 and with those structures cited in the literature (Dall’Acqua and Innocenti, 2007; Pawar et al., 2007). The MS spectra of the less abundant peaks observed in the chromatogram showed carbohydrate neutral losses and steroid backbone ions

consistent with those seen for H.g.-12 and the other named steroid glycosides demonstrating structural similarity. The linear un-weighted calibration gave correlation >0.999. Blank chromatograms showed a small peak (<5 mAU) eluting at ca 46 min which was excluded from the steroid glycoside quantitation. The seven steroid glycosides, along with the aglycone, were quantified against the H.g.-12 standard line and the relative response factors reported previously (Janssen et al., 2008) applied to correct for differing molecular masses of each species. Assay reproducibility was excellent as demonstrated by the low (<1)%RSD figures (Table 1) for the named steroid glycosides with the exception of the low level aglycone, which showed slightly greater variability. The species reported as demonstrating satiety activity (van Heerden et al., 2007), H.g.-12 was present in the extract at a level of 10.2% w/w (±0.3% RSD) with the seven named steroid glycosides totalling 33.2% w/w. Other steroid glycosides were defined as those peaks not present in the blank that eluted after 15 min and accounted for 30.5% w/w resulting in a total steroid glycoside content of 73.7% w/w (±0.5% RSD) in the H. gordonii extract. Subsequent batches of H. gordonii extracts produced by the same solvent extraction procedure as the material reported here have demonstrated consistency with the variation in H.g.-12 and total steroid glycoside content not varying by greater than ±10%. The LC–UV/MS fingerprint profiles of these subsequent batches have shown excellent similarity demonstrating robustness of the extraction procedure. The major components in the saponified sample were identified by GC–MS using the results of the C4–C24 FAME standard in conjunction with NIST_05 spectral library and transfer to the quantitative GC-FID data was done by visual peak recognition and retention time comparison. The area sum of all peaks, not present in the blank, in the GC-FID chromatogram was quantified using the area and known weight of the C11:0 fatty acid internal standard to give the total GC-FID fraction (% w/w). Cholesterol, b-sitosterol, atocopherol and stigmasterol were the main components present in this fraction at a total level of 0.39% w/w (±12.1% RSD). In addition to these compound classes the GC-FID chromatogram showed a large number of minor unidentified peaks which contributed 3.09% w/w (±5.6% RSD) to the total quantitation. Mass balance by individual component analysis was >80% which was considered acceptable given the high sample complexity. Ash content did not significantly add to this (<1% w/w) and it is speculated that the balance could be increased by further investigation into the ‘other steroid glycosides’ which have a mean UV response factor applied to them. The volume of water/acetonitrile/methanol (50/35/15 v/v/v) required to elute the polar material from the SPE column was variable between replicates (5–20 ml). An aliquot of the eluent collected was analysed on the LC–UV/ELSD system, showing only early eluting compounds to be present. However, analysis of later fractions indicated that all of the early eluting material had not been isolated and some was continuing to elute with the steroid glycosides. Consequently the results are quoted as a minimum figure (3.1% w/w), but the contribution of non-isolated polar material was considered to be minimal (<2% w/w). After elution and analysis of the polar material from the SPE cartridge, a final elution of 50 ml of methanol was carried out to elute steroid glycosides and to quantify any mass losses that may have occurred during the separation process. The mass balance of recovered material (polar organics, steroid glycosides and filter residues) obtained by this process totaled >90%. Protein analysis conducted using SDS–PAGE gave no evidence of the presence of protein (detection limit of <0.001% w/w) which was supported by Kjeldahl nitrogen analysis results (not reported). These results indicated that the potential for H. gordonii extract to illicit an allergenic response is minimal.

P.J. Russell, C. Swindells / Food and Chemical Toxicology 50 (2012) S6–S13

S11

Fig. 1. HPLC–UV (220 nm) Chromatogram of Hoodia gordonii extract.

H. gordonii extract gave a satisfactory contaminant profile (Table 2) to allow a controlled population exposure within toxicological and clinical studies. For such novel materials there are no specific regulatory limits; however, the lowest specification or that of the

closest matched material has been considered. There was sufficient evidence obtained from the different production batches characterised to confirm that a commercial scale food grade extract was obtainable using the extraction process reported previously.

Fig. 2. MS Fragmentation pattern for H.g.-12.

S12

P.J. Russell, C. Swindells / Food and Chemical Toxicology 50 (2012) S6–S13

Fig. 3. Chemical structures of key steroid glycosides.

P.J. Russell, C. Swindells / Food and Chemical Toxicology 50 (2012) S6–S13

6. Conclusion H. gordonii extract has been extensively characterised with a large number of individual components identified. Total steroid glycoside levels were >70% w/w with H.g.-12 quantified at 10.2% w/w. Contaminant screening and protein analysis demonstrated that a safe commercial scale extract could be produced which has a good stability profile.

Conflict of Interest The authors declare that there are no conflicts of interest. Acknowledgements The following colleagues within Unilever are thanked for their contribution to the analytical study of Hoodia gordonii extract: Romain Descamps, Hans-Gerd Janssen, Anja Lalljie, Juliette Pickles Ian Sanders and Liz Tulum. References Avula, B., Wang, Y.H., Pawar, R.S., Shukla, Y.J., Schaneberg, B., Khan, I.A., 2006. Determination of the appetite suppressant P57 in Hoodia gordonii plant extracts and dietary supplements by liquid chromatography/electrospray ionization mass spectrometry (LC-MSD-TOF) and LC–UV methods. Journal of AOAC International 89 (3), 606–611. Bruyns, P.V., 2005. Stapeliads Africa and Madagascar. Umdaus Press. ISBN 1919766-38-3.

S13

Cheetham, S.C., Jackson, H.C., Vickers, S.P., Dickinson, K., Jones, R.B., Heal, D.J., 2004. Novel targets for the treatment of obesity: a review of progress. Drug Discovery Today: Therapeutic Strategies 1 (2), 227–235. Dall’Acqua, S., Innocenti, G., 2007. Steroidal glycosides from Hoodia gordonii. Steroids 72 (6–7), 559–568. Dent, M. P., Wolterbeek, A.P.M., Russell P. J., Bradford, R. (2012a) Safety profile of Hoodia gordonii extract: Mouse prenatal developmental toxicity study Food and Chemical Toxicology 50 supplement 1 S20–S25. Published online ahead of print doi:10.1016/j.fct.2011.06.017. Dent, M.P., Wolterbeek, A.P.M., Russell P.J., Bradford, R. Safety profile of Hoodia gordonii extract: Rabbit prenatal developmental toxicity study (2012b) Food and Chemical Toxicology 50 supplement 1 S26–S33. Published online ahead of print doi:10.1016/j.fct.2011.04.008. Harrold, J., Pinkney, J., Williams, G., 2004. Control of obesity through the regulation of appetite. Drug Discovery Today 1 (2), 219–225. Janssen, H.G., Swindells, C., Gunning, P., Wang, W., Grun, C., Mahabir, K., Maharaj, V.J., Apps, P.J., 2008. Quantification of appetite suppressing steroid glycosides from Hoodia gordonii in dried plant material, purified extracts and food products using HPLC–UV and HPLC–MS methods. Analytica Chimica Acta 617 (1–2), 200–207. Knight, T.L., Swindells, C.M., Craddock, A.M., Maharaj, V.J., Buchwald-Werner, S., Ismaili S.A., McWilliam, S.C. (2012) Cultivation practices and manufacturing processes to produce Hoodia gordonii extract for weight management products. Food and Chemical Toxicology 50 supplement 1 S1–S5. Published online ahead of print doi:10.1016/j.fct.2011.09.042. Pawar, R.S., Shukla, Y.J., Khan, I.A., 2007a. New calogenin glycosides from Hoodia gordonii. Steroids 72 (13), 881–891. Pawar, R.S., Shukla, Y.J., Khan, S.I., Avula, B., Khan, I.A., 2007b. New oxypregnane glycosides from appetite suppressant herbal supplement Hoodia gordonii. Steroids 72 (6–7), 524–534. Scholz, E., 1984. Karl-Fischer-Titration. Springer-Verlag. Unilever R&D Vlaardingen (Internal Report) 2008, Interlaboratory comparison measurement of Hoodia gordonii actives in oven dried plant material and purified extracts. van Heerden, F.R., Marthinus, H.R., Maharaj, V.J., Vleggaar, R., Senabe, J.V., Gunning, P.J., 2007. An appetite suppressant from Hoodia species. Phytochemistry 68 (20), 2545–2553.