J Sci Food Agric 1998, 78, 141È146 Effect of Storage and Extraction on Ratio of Soyasaponin I to 2,3-Dihydro-2,5-dihydroxy-6methyl-4-pyrone-Conjugated Soyasaponin I in Dehulled Peas (Pisum sativum L) Ylva Dandanell Daveby,1* Per A man,1 Joseph M Betz2 and Steven M Musser2 1 Department of Food Science, Swedish University of Agricultural Sciences, S-750 07 Uppsala, Sweden 2 US Food and Drug Administration, Center for Food Safety and Applied Nutrition, HFS-347, 200 C St SW, Washington, DC 20204, USA (Received 5 March 1997 ; revised version received 4 August 1997 ; accepted 20 October 1997) Abstract : The e†ects of solvent and time on extraction of 2,3-dihydro-2,5dihydroxy-6-methyl-4-pyrone (DDMP)-conjugated soyasaponin I in dehulled peas (Pisum sativum L) were studied. Extraction in 80% ethanol gave a higher and more stable yield than 100% methanol over the time studied. Acetonitrile did not extract any detectable saponins. Both soyasaponin I and the DDMPconjugated form were present in the extracts of dehulled, dried peas. The proportion of soyasaponin I to the DDMP-conjugate, calculated as the ratio of the molecular ions, increased with increasing extraction time, from an initial 0É24 after 0É5 h to 0É55 after 24 h of extraction in 80% ethanol, showing that the DDMP-conjugate was hydrolysed during the extraction. Storage of dried dehulled peas also increased the proportion of soyasaponin I, from 0É063 in newly harvested peas to 0É34 after 9 months of storage. Peas harvested in 1988 and stored for 7 years had a ratio of 0É78. These results clearly show that the DDMPconjugate was converted to soyasaponin I both during storage and extraction of the peas. ( 1998 Society of Chemical Industry. J Sci Food Agric 78, 141È146 (1998) Key words : Pisum sativum, dehulled seeds, hulls, soyasaponins, DDMPconjugate, storage, extraction. Phaseolus coccineus (Yoshiki et al 1994), L upinus angustifolius (Ruiz et al 1995), Medicago sativa and Glycine max (Massiot et al 1992) are soyasaponins conjugated with 2,3-dihydro-2,5-dihydroxy-6-methyl-4-pyrone(DDMP) (Fig 1). Traditional methods of extracting, derivatising and analysing soyasaponins have used hot exhaustive extractions. It has been suggested that heating converts DDMP-conjugated soyasaponin I to soyasaponin I (Fig 2) (Kudou et al 1992). Later studies on soyasaponins in legumes using mild (eg room temperature) extractions have shown DDMPconjugated soyasaponins to be the predominant soyasaponins (Kudou et al 1992 ; Massiot et al 1992). The DDMP-conjugated soyasaponins have a strong superoxide anion (O~) scavenging activity, suggesting a pre2 ventive role against biomolecular damage due to radical INTRODUCTION Saponins are steroid or triterpenoid glycosides present in a wide variety of plants consumed by humans and animals. The soyasaponins comprise a triterpenoid soyasapogenol with one or two (mono- or bidesmosidic) carbohydrate sidechains attached. They have a bitter taste (Price and Fenwick 1984), inhibitory e†ects against infectivity of the AIDS virus (Okubo et al 1994) as well as possible cholesterol-binding and growth retarding activities (Cheeke 1996). These properties are, however, dependant upon the individual structure of the soyasaponins (Okubo et al 1994). It has been reported that the native saponins in Pisum sativum (Tsurumi et al 1992), * To whom correspondence should be addressed. 141 ( 1998 Society of Chemical Industry. J Sci Food Agric 0022È5142/98/$17.50. Printed in Great Britain Y D Daveby et al 142 Fig 1. MS/MS product ion spectrum of m/z 1069 (DDMP-conjugate soyasaponin I). attack (Yoshiki and Okubo 1995). It has also been suggested to function as a reductant in root growth regulation (Tsurumi and Tsujino 1995). In peas, a domestic protein crop in Sweden, only soyasaponin I has been quantiÐed (Curl et al 1985 ; Price et al 1986 ; Daveby et al 1997). DDMP-conjugated soyasaponin I was quantiÐed in 7-day-old aetiolated pea seedlings (Tsurumi et al 1992). The latter study reported production of soyasaponin I from DDMPconjugated soyasaponin I in frozen tissues or from prolonged extraction. We have investigated the efficiency of di†erent solvents for extraction of DDMP-conjugated soyasaponin I as well as the e†ect of storage and extraction time on the ratio of soyasaponin I to DDMP-conjugated soyasaponin I in dehulled peas. MATERIALS AND METHODS Materials Swedish light-coloured and leaÑess peas (Pisum sativum L ; subsp hortense ; cv Capella) grown during 1988 at Ultuna, Sweden (60¡N) and during 1995 at Svalov, Sweden (56¡N) were studied. The peas were soaked in distilled water at 4¡C overnight, dehulled by hand (the germs were included with the dehulled seeds) and the dehulled seeds were freezedried. The peas harvested in 1988 were stored dry as whole seeds in room temperature until analysis. Peas harvested in 1995 were stored dehulled in a desiccator until analysis. Prior to analysis samples were ground in a Tecator cyclone sample mill to pass a 0É5 mm sieve. One batch each from the 1995 (in a desiccator) and the 1988 harvest were stored whole and ground (0É5 mm) prior to analysis. All results are based on duplicate analysis with \5% and \8% as highest acceptable differences between duplicate samples for HPLC analyses and LC-MS calculated ratios, respectively. Extraction To investigate the extraction efficiency, the dehulled peas (0É5 g) were extracted in a shaker at room temperature for 2, 4, 6, 8 and 10 h in 80% aqueous ethanol, 100% acetonitrile or 100% methanol (5 ml). The samples were centrifuged (875 ]g, 10 min) and Ðltered (0É45 lm) prior to analysis by HPLC. Ratios of soyasaponin I to DDMP-conjugated soyasaponin I The ratios of the area of the molecular ion of soyasa- E†ect of storage and extraction on soyasaponins 143 Fig 2. MS/MS product ion spectrum of m/z 943 (soyasaponin I). ponin I ([MH]` \ 943) to the corresponding area of the DDMP-conjugated soyasaponin I ([MH]` \ 1069) were analysed using LC-MS after extraction of the dehulled peas at room temperature in 80% ethanol for 0É5, 2, 4, 6, 12, 18 and 24 h. The ratios were also analysed after 0É5 h extraction in 80% ethanol, followed by Ðltration and standing at room temperature for 1É5, 5É5, 11É5, 17É5 and 23É5 h prior to analysis. HPLC Saponins were analysed by HPLC on a YMC (YMC Inc., Wilmington, NC, USA) J-sphere ODS-L80 S-4 lm (4É6 ] 250 mm) steel cartridge. After initial scanning with a photodiode array detector, detection was carried out at 205 and 292 nm. The following equipment was used : Waters 600 E Multisolvent Delivery System (Waters Associates, Milford, CA, USA, LDC spectromonitor 3100 (LDC Analytical, Riviera Beach, FL, USA), Waters 991 Photodiode Array Detector, BioRad AS-100 autosampler (BioRad Inc, Hercules, CA, USA) and SP 4270 integrator (Spectra Physics, San Jose, CA, USA). The column was eluted at a Ñow rate of 1 ml min~1 starting with 35% CH CN followed by a 3 linear gradient to 60% CH CN in 40 min ; a concentra3 tion of 40 mM acetic acid was maintained throughout elution. For post-column collection for MS/MS experiments the 80% ethanol extract was eluted through a YMC Inc. J-sphere ODS-L80 S-4 lm (10 ] 250 mm) steel cartridge with a Ñow rate of 2É4 ml min~1. All other conditions were the same as above. The fractions (26É71 and 27É37 min) were collected over dry ice and immediately stored in a freezer ([30¡C). Prior to analysis the acetonitrile was evaporated under nitrogen, and the sample lyophilised. LC-MS experimental A Hewlett-Packard Model 1050 LC pump (Palo Alto, CA, USA) was used to provide linear gradients and a constant Ñow rate of 200 ll min~1. All chromatography was performed on a YMC Inc. J-sphere ODS-L80 LC column (2 ] 250 mm). Chromatographic elution for positive ion electrospray analysis started with 18% CH CN followed by a linear gradient to 81% CH CN 3 3 in 30 min ; a concentration of 40 mM formic acid was maintained throughout elution. Under these conditions soyasaponin I elutes at 20É0 min and DDMPconjugated soyasaponin I at 22É1 min. A Finnigan Y D Daveby et al 144 Model TSQ-7000 triple-quadrupole mass spectrometer with the standard Finnigan electrospray ion source was used for ionisation of the analytes. Nitrogen was used as a nebulising gas and the capillary temperature was 225¡C. The instrument was scanned over the range of 600È1400 amu at 1 s per scan. The entire 200 ll min~1 column effluent was directed into the ion source. For MS/MS experiments argon was used as the collision gas along with a collision energy of 35 V. RESULTS AND DISCUSSION Attempts were made to isolate a standard of DDMPconjugated soyasaponin I. Owing to degradation during the isolation procedure we were not able to isolate a pure compound which could be used as a standard. Chromatographic runs with a photodiode array detector were compared for a standard of soyasaponin I and the ethanol extract. The soyasaponin I eluted at 26É71 min with a maximum absorbance at 205 nm. The ethanol extract showed two distinct and well-separated peaks at 26É71 and 27É37 min, the latter with maximum absorption at 292 nm and the former with smaller absorption at 205 nm. These data are in very good agreement with earlier reports for soyasaponin I and DDMP-conjugated soyasaponin I (Kudou et al 1992 ; Massiot et al 1992 ; Yoshiki et al 1995). MS/MS product ion spectrum of m/z 943 from the 26É71 min collected fraction gave a molecular ion at m/z 943 and fragment ions at 797, 635, 459, 441 and 423 (Fig 2). These are expected fragment ions for soyasaponin I and veriÐes the identify of this peak as soyasaponin I. MS/MS product ion spectrum of the 27É37 min collected fraction gave a molecular ion at m/z 1069 and fragment ions at 923, 759, 581, 567, 423, 144 and 126 (Fig 1). These data are in very good agreement with earlier reports on DDMP-conjugated soyasaponin I (Kudou et al 1993 ; Yoshiki et al 1994). Together with the chromatographic appearance the compound was identiÐed as DDMP-conjugated soyasaponin I. The yield of DDMP-conjugated soyasaponin I (in relative area units) from the extraction procedures are shown in Fig 3. Extracting with 80% ethanol gave the highest and most stable recovery over the time studied. The amount of saponin extracted increased up to about 4 h. Extraction times from 4 to 32 h gave similar yields. Extracting with 100% methanol gave a very low yield of saponins. One hundred percent acetonitrile did not extract any detectable amounts of DDMP-conjugated saponins. Both soyasaponin I and DDMP-conjugated soyasaponin I were found in the extracts of mature dehulled peas. However, the ratio of the two saponins changed during storage and with di†erent extraction times. In newly harvested mature dried dehulled peas the ratio of Fig 3. Yield of DDMP-conjugated soyasaponin I with di†erent extraction solvents. =, Methanol ; K, 80% ethanol. the areas of the molecular ions of 943 : 1069 after 0É5 h extraction was 0É063. After storage of the dehulled peas in a desiccator for 5 and 7 months, the ratio had increased to 0É19 and 0É24, respectively. When stored whole in a desiccator and ground without dehulling after 9 months the ratio was 0É34. The same cultivar of peas harvested in 1988, stored as whole seeds at room temperature for 7 years and dehulled and ground prior to analysis, had a ratio of 0É76. In peas from the 1988 harvest, stored at room temperature for 7 years and analysed without dehulling, the ratio was 0É78. The proportion of soyasaponin I to the conjugated form, also increased with increasing extraction time, from an initial ratio of 0É24 after 0É5 h in 80% ethanol to 0É55 after 24 h of extraction (Fig 4). Exclusion of light during extraction had no e†ect. When extracting for 0É5 h, Ðltering and then standing at room temperature prior to analysis, the ratio of the areas of the molecular ions increased in a similar manner but at a lower initial rate. These results suggest continued release of DDMP-conjugated soyasaponin I from the pea extract up to approximately 4 h. Prolonged extraction in 80% ethanol seems to enhance the conversion of DDMP-conjugated soyasaponin I to soyasaponin I, the rate after about 4 h being similar with or without the ground peas in the extract. The half life of DDMP-conjugated soyasaponin I (approximately 150 h) was calculated from the results for the breakdown of DDMP-conjugated soyasaponin I to soyasaponin I at room temperature after 0É5 h extraction. The ratio of [MH]` \ 943 : [MH]` \ 1069 during the extraction may be determined by the rate of extraction for the compounds, half life of DDMPconjugated soyasaponin I and the relative composition in the starting material. The Ðtted line for the ratio of E†ect of storage and extraction on soyasaponins 145 [MH]` \ 943 : [MH]` \ 1069 after storage. With the ratio in whole peas being similar to the ratio in dehulled peas, no e†ect of soaking, dehulling or freeze-drying was found. Neither did exposure to light a†ect the ratio compared to extraction in darkness. It is thus suggested that not only heating or prolonged extraction but also storage after harvesting releases soyasaponin I from the DDMP-conjugate form, which could be due to natural enzymatic processes in the cotyledon. Further evaluation of the compositional changes of soyasaponins during storage and extraction is needed. ACKNOWLEDGEMENTS Fig 4. E†ect of extraction time on ratio of [MH]` \ 943 : [MH]` \ 1069. =, Extracted in 80% ethanol for 0É5 h and then standing in 80% ethanol at room temperature ; K, extraction with 80% ethanol. The Ðtted line represents the ratio obtained by simulation of the course of extraction. [MH]` \ 943 : [MH]` \ 1069 after extraction in 80% ethanol in Fig 4 represents the ratio obtained by simulation of the course of extraction using the previously calculated half-life as well as the optimised rate of extraction and starting ratio. Earlier results from Glycine max seeds (Kudou et al 1993), Phaseolus coccineus hypocotyls (Yoshiki et al 1994) and L upinus angustifolius seeds (Ruiz et al 1995) showed only the DDMP-conjugated soyasaponins to be present after mild extraction. In 7-day-old aetiolated pea seedlings only DDMP-conjugated soyasaponin I was reported and soyasaponin I was not found in the free form. The concentration of DDMP-conjugate was higher in the hook and root tip than in other nongrowing tissues (Tsurumi et al 1992). This study also reported release of soyasaponin I from the DDMPconjugated form when extracting frozen tissues or with prolonged extraction of unfrozen tissues. Soybean hypocotyls extracted for 30 min at room temperature had DDMP-conjugated saponins I and II as major compounds whereas after extraction for 5 h at 80¡C the major compounds were soyasaponin I and II (Kudou et al 1992). Heating of DDMP-conjugated soyasaponins at 100¡C for 1 h completely converted them into the corresponding soyasaponins (Kudou et al 1994). DDMP-conjugated soyasaponin I was the only saponin detected in unprocessed chickpeas and lentils. Soaking of the seeds for 24 h at 25¡C at di†erent pH values did not a†ect the saponin composition (Ruiz et al 1996). The present study indicates that both soyasaponin I and the DDMP-conjugate are present in the mature dehulled peas, with an increasing ratio of the areas of The authors are indebted to those at the US Food and Drug Administration who provided the opportunity to perform this work. Special thanks to Dr Samuel W Page for placing the resources of his research group at the authors disposal and to Dr James A Sphon for allowing access to the LC-MS instrumentation used in these studies. REFERENCES Cheeke P R 1996 Biological e†ects of feed and forage saponins and their impacts on animal production. In : Saponins used in Food and Agriculture, ed Waller G R & Yamasaki K. Plenum Press, New York, USA, pp 377È385. Curl C L, Price K R, Fenwick G R 1985 The quantitative estimation of saponin in pea (Pisum sativum L.) and soya (Glycine max). Food Chem 18 241È250. Daveby Y D, Óman P, Betz J M, Musser S M, Obermeyer W R 1997 The variation in content and changes during development of soyasaponin I in dehulled Swedish peas (Pisum sativum L). J Sci Food Agric (In press). Kudou S, Tonomura M, Tsukamoto C, Shimoyamada M, Uchida T, Okubo K 1992 Isolation and structural elucidation of the major genuine soybean saponin. Biosci Biotechnol Biochem 56 142È143. Kudou S, Tonomura M, Tsukamoto C, Uchida T, Sakabe T, Tamura N, Okubo K 1993 Isolation and structural elucidation of DDMP-conjugated soyasaponins as genuine saponins from soybean seeds. Biosci Biotechnol Biochem 57 546È550. Kudou S, Tonomura M, Tsukamoto C, Uchida T, Yoshikoshi M, Okubo K 1994 Structural elucidation and physiological properties of genuine soybean saponins. In : Food Phytochemicals for Cancer Prevention I. Fruit and V egetables, ed Huang M-T et al . American Chemical Society, Washington DC, USA, pp 340È348. Massiot G, Lavaud C, Benkhaled M, Le Men-Olivier L 1992 Soyasaponin VI, a new maltol conjugate from alfalfa and soybean. J Nat Prod 55 1339È1342. Okubo K, Iudou S, Uchida T, Yoshiki Y, Yoshikoshi M, Tonomura M 1994 Soybean saponin and isoÑavonoids. In : Food Phytochemicals for Cancer Prevention I. Fruit and V egetables, ed Huang M-T et al. American Chemical Society, Washington DC, USA, 330È339. 146 Price K R, Curl C L, Fenwick G R 1986 The saponin content and sapogenol composition of the seed of 13 varieties of legume. J Sci Food Agric 37 1185È1191. Price K R, Fenwick G R 1984 Soyasaponin I, a compound possessing undesirable taste characteristics isolated from the dried pea (Pisum sativum L) J Sci Food Agric 35 887È 892. Ruiz R G, Price K R, Rose M E, Rhodes M J C, Fenwick G R 1995 Determination of saponins in lupin seed (L upinus angustifolius) using high-performance liquid chromatography : comparison with a gas chromatographic method. J L iquid Chromatogr 18 2843È2853. Ruiz R G, Price K R, Arthur A E, Rose M E, Rhodes M J, Fenwick R G 1996 E†ect of soaking and cooking on the saponin content and composition of chickpeas (Cicer arietinum) and lentils (L ens culinaris). J Agric Food Chem 44 1526È1530. Y D Daveby et al Tsurumi S, Tsujino Y 1995 Chromosaponin I stimulates the growth of lettuce roots. Physiologica Plantarum 93 785È789. Tsurumi S, Takagi T, Hashimoto T 1992 A y-pyronyl-triterpenoid saponin from Pisum sativum. Phytochemistry 31 2435È2438. Yoshiki Y, Kim J-H, Okubo K 1994 Saponins conjugated with 2,3-dihydro-2,5-dihydroxy-6-methyl-4H-pyran-4-one from Phaseolus coccineus. Phytochemistry 36 1009È1012. Yoshiki Y, Kim J H, Okubo K, Nagoya I, Sakabe T, Tamura N 1995 A saponin conjugated with 2,3-dihydro-2,5dihydroxy-6-methyl-4H-pyran-4-one from Dolichos lablab. Phytochemistry 38 229È231. Yoshiki Y, Okubo K 1995 Active oxygen scavenging activity of DDMP (2, 3-dihydro-2,5-dihydroxy-6-methyl-4H-pyran4-one) saponin in soybean seed. Biosci Biotechnol Biochem 59 1556È1557.