Extractability of Thevetia peruviana glycosides with alcohol mixture

The aim of this work was to investigate the most effective extraction conditions for the production of thevetia seed protein concentrate of reduced cardiac glycoside content. Alcoholic extraction of the glycosides was studied as a function of time, solvent to meal ratio and solvent composition. Thevetia seed meal was extracted with 10:1, 15:1 and 20:1 solvent to meal ratios, for 45 min, 24, 48 and 72 h. Varying concentrations - 50 to 100% (v/v) aqueous alcohol - were also used. A concentration of 70 or 80% aqueous alcohol resulted in the lowest glycoside content, while a solvent to meal ratio of 15:1, extracted over a period of 72 h, appears to give the best compromise between glycoside extraction and cost of extraction solvent. All treatments resulted in an increase in the protein content of the samples.

Ingestion of thevetia seed by either man or animal has been reported to result in severe cardiac toxicity which produced marked poisoning symptoms that culminated in death (Oji and Okafor, 2000;de Silva et al., 2003;Fonseka et al., 2002).Symptoms include nausea, vomting and giddiness within hours.Other clinical features are severe diarrhoea, abdominal pain, dilated pupils and occasionally convulsions (Fernando and Widyaratna, 1989;Samal et al., 1992).
A detoxification of this seed will be required before the seed can find application as a possible protein source in animal feed formulation.Several attempts have been *Corresponding author.E-mail: laraoluwaniyi@yahoo.com made in the past to detoxify the seed cake with little success.Some of the methods that have previously been employed include dry heating, autoclaving and fermentation (Odetokun et al., 1999;Taiwo et al., 2004).These methods did not record much success in the reduction of the cardiac glycosides of the seed.Finnigan and Lewis (1985) have shown that oilseed protein quality may be improved by alcoholic extraction of undesirable components to leave a protein concentrate.Van Megan (1983) has also reported ethanolic extraction of defatted rapeseed to be successful in detoxification.We also recently reported the reduction in the glycoside content of thevetia following alcoholic extraction of the defatted meal (Oluwaniyi et al., 2007).This current study therefore aims at investigating the most favourable conditions for the alcoholic extraction of the glycosides of thevetia seed meal with respect to the time taken for extraction, quantity and composition of extractant used, reduction in the content of the glycosides as well as the quantity of protein in the residual meal.

Materials
Matured fruits of thevetia plant were collected from a location in Ilorin, Kwara State, Nigeria, by direct plucking of matured (black) fruits from plants and by picking those that had fallen off plants.The  fruits were cracked to remove the hard pericarp and mesocarp and the soft seeds crushed into a paste.The paste was defatted first by mechanical pressing, followed by solvent extraction using predistilled n-hexane.

General extraction methods
Alcoholic extraction of thevetia seed meal was studied as a function of time, solvent to meal ratio and solvent composition as described by Finnigan and Lewis (1988).All extractions were performed at ambient temperature.A sample of thevetia seed meal was placed in a flask and solvent was added to give the appropriate solvent to meal ratio.Extraction at 45 min was achieved using a magnetic stirrer while the others were covered and left.Alcoholic extraction was performed using aqueous 8:2 ethanol:methanol mixture.Solvent to meal ratios of 10:1, 15:1 and 20:1 were investigated.
Effect of time and solvent to meal ratio 10 g each of thevetia seed meal was extracted with 80% (v/v) aqueous alcoholic mixture at solvent to meal ratios of 10:1.15:1 and 20:1.The extractability of each solvent to meal ratio was also investigated as a function of time.Each was extracted for 45 min, 24, 48 and 72 h.At the end of the experimental time, the samples in the flasks were filtered, the solvent squeezed out and the residue air-dried.

Effect of solvent composition
Thevetia seed meal was extracted with 50, 60, 70, 80, 90 and 100% (v/v) aqueous alcohol mixture at a solvent to meal ratio of 15: 1 for 72 h after which each was filtered, squeezed and dried.

Moisture and protein content determination
The moisture content of each sample (treated and untreated) was determined by heating 2.0 g of each sample to a constant weight in a crucible placed in an oven maintained at 105ºC (AOAC, 1984).
The nitrogen content was determined by the Kjeldahl digestion method (Kjeldahl, 1883), using 2.0 g samples.The crude protein was calculated by multiplying the nitrogen content with a factor of 6.25.

Glycoside content determination
The quantity of cardiac glycosides in the raw and treated samples was evaluated using Baljet's reagent (95 ml aqueous picric acid + 5 ml 10% aqueous NaOH) as described by El-Olemy et al. (1994). 1 g of each sample was soaked in 10 ml of 70% alcohol for 2 h and then filtered.The extracts obtained were then purified using lead acetate and Na2HPO4 solutions before the addition of freshly prepared Baljet's reagent.The intensity (absorbance) of the colour produced was then measured using a spectrophotometer at 495 nm.The difference between the intensity of the colours of the experimental and blank (distilled water and Baljet's reagent) samples gives the absorbance and is proportional to the concentration of the glycoside.

RESULTS AND DISCUSSION
Analysis of thevetia seed meal showed that the untreated meal contains 11.83% moisture, 54.4 g/kg (or 5.44 g %) total cardiac glycosides and 542.5 g/kg (54.25%) of crude protein.
The result of the effect of time and solvent to meal ratio on the extractability of the toxins of thevetia meal (which are mainly cardiac glycosides) is illustrated in Figure 2. The glycoside content was reduced by 91.5% with 10: 1 alcohol to thevetia seed meal, 94.1% with 15:1 and by 95.6% with 20:1 alcohol to meal after 72 h.The protein content as presented in Table 1 also shows an increase in the protein content of the residual protein meal after treatment with the different quantities of solvent when compared with the untreated meal.Extracting with 15:1 solvent to meal ratio however appears to give the best result as far as conserving the protein is concerned.The protein contents ranged between 68.70-69.09%,whereas treating with either a higher or lower proportion of solvent resulted in a decrease in the protein content.The length of time used for treating the meal does not seem to have an effect on the protein content of the product.
Figure 3 shows the effect of varying the concentration of aqueous alcohol used in the extraction of the cardiac glycosides.The results show that extracting with 70-90% aqueous alcohol gave the meal with the low glycolside content whereas using 100% alcohol resulted in the poorest extraction of the glycosides.Preliminary work done in our laboratory also reveal that there is little difference (if any), between the extraction recorded with ethanol and that obtained with the ethanol/methanol mixture.
The effect of varying the solvent composition on the protein content (Table 2) also follows a similar trend with the effect on glycoside content.Extraction with 80 and 70% aqueous alcohol resulted in the highest protein content in the residual meal while the least protein content was obtained from the meal extracted with 50 and 100% aqueous alcohol.It was also observed that the protein concentrate remaining after treatment with 50% alcohol was very susceptible to spoilage, which is attributed to the high percentage of water in the solvent, thus confirming further that a solvent composition that is less than 70% or greater than 90% alcohol results in a poor extraction of the glycoside as well as a loss of the crude protein.
The final products, irrespective of the concentration and ratio of solvent or duration of treatment, still contain a residual amount of cardiac glycoside although quite minute.It is however possible that double extraction of the meal may result in a meal with undetectable cardiac glycoside content.
These results show that the cardiac glycosides of thevetia seed meal can be significantly reduced and this

Figure 1 .
Figure 1.Structures of cardiac glycosides and their free aglycones.

Table 1 .
Moisture and protein contents (%) of treated thevetia seed meal as a function of time and solvent to meal ratios.

Table 2 .
Effect of varying concentrations of aqueous alcohol (15:1 solvent to meal ratio) on the moisture and protein contents (%) of treated thevetia seed meal.an increase in the protein content of the residual meal.thevetia seed meal treated this way can therefore serve as a protein source for animal feed formulation.