Biguanide related compounds in traditional antidiabetic functional foods

Food Chemistry 138 (2013) 1574–1580

Contents lists available at SciVerse ScienceDirect

Food Chemistry journal homepage: www.elsevier.com/locate/foodchem

Biguanide related compounds in traditional antidiabetic functional foods Venu Perla, Sastry S. Jayanty ⇑ San Luis Valley Research Center, Department of Horticulture and Landscape Architecture, Colorado State University, 0249 East County Road 9N, Center, CO 81125, USA

a r t i c l e

i n f o

Article history: Received 10 April 2012 Received in revised form 23 July 2012 Accepted 21 September 2012 Available online 23 November 2012 Keywords: Biguanide Arginine Urea Antidiabetic Functional food Ayurveda

a b s t r a c t Biguanides such as metformin are widely used worldwide for the treatment of type-2 diabetes. The identification of guanidine and related compounds in French lilac plant (Galega officinalis L.) led to the development of biguanides. Despite of their plant origin, biguanides have not been reported in plants. The objective of this study was to quantify biguanide related compounds (BRCs) in experimentally or clinically substantiated antidiabetic functional plant foods and potatoes. The corrected results of the Voges–Proskauer (V–P) assay suggest that the highest amounts of BRCs are present in green curry leaves (Murraya koenigii (L.) Sprengel) followed by fenugreek seeds (Trigonella foenum-graecum L.), green bitter gourd (Momordica charantia Descourt.), and potato (Solanum tuberosum L.). Whereas, garlic (Allium sativum L.), and sweet potato (Ipomea batatas (L.) Lam.) contain negligible amounts of BRCs. In addition, the possible biosynthetic routes of biguanide in these plant foods are discussed. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction Type-2 diabetes is the most common metabolic disorder worldwide, and its prevalence is growing at alarming rates in both developed and developing countries. Oral antidiabetic agents commonly available on the market to treat type-2 diabetes include biguanides, sulphonylureas, thiazolidinediones, meglitinides and a-glucosidase inhibitors (Mahajan & Gupta, 2010). Among biguanides, metformin (1,1-dimethylbiguanide), also known as ‘glucophage’, is the most widely prescribed drug used in the treatment of diabetes. The history of metformin can be traced back to the use of French lilac (goat’s rue, Spanish sanfoin, or false indigo) as herbal medicine in medieval Europe. French lilac is rich in guanidine, a substance with blood glucose-lowering (hypoglycaemic) activity that is present in the basic structure of metformin (Bailey & Day, 2004) (Fig. 1). Similarly, indigenous herbal medicines have been used in the treatment of diabetes in Ayurvedic medicine practiced in India by Charaka and Sushruta since 6th century BC (Grover, Yadav, & Vats, 2002). Unlike guanidine, biguanides (such as metformin) are non-toxic to animals and safe for clinical use in diabetic patients (Bailey & Day, 2004). Because biguanides do not increase pancreatic insulin secretion, they are referred to as antihyperglycaemic agents as opposed to hypoglycaemic agents (sulphonylureas). Biguanides reduce hyperglycaemia by increasing insulin sensitivity, decreasing glucose absorption, and inhibiting hepatic gluconeogenesis. Metformin is currently available in the drug mar⇑ Corresponding author. Tel.: +1 7197543594; fax: +1 7197542619. E-mail addresses: [email protected] (V. Perla), [email protected] (S.S. Jayanty). 0308-8146/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.foodchem.2012.09.125

ket because of its unique mechanism of action, lower risk of lactic acidosis as compared to phenformin, and its successful use in over 90 countries. Suppression of hepatic glucose production and increased peripheral insulin sensitivity appear to be the major mechanisms of action by which metformin restores glycaemic control (Goo, Carson, & Bjelajac, 1996). In addition, biguanides are also used as antimalarial drugs. Furthermore, these compounds are known to have oral hypoglycaemic, tumor-inhibiting, antibacterial, tuberculostatic and antiviral properties (Kurzer & Pitchfork, 1968). The less toxic guanidine-containing compound, galegine is also an antidiabetic compound isolated from French lilac (Bailey & Day, 2004). Galegine is known for its weight-reducing properties indirectly by inhibiting synthesis and stimulating oxidation of fatty acids (Mooney et al., 2008). L-Arginine is a guanidine-containing amino acid known to stimulate the release of hormones, such as insulin, glucagone, prolactine and growth hormone (Tousoulis et al., 2002). Interestingly, the structure of guanidine is similar to that of urea (Barritt, 1936). Biuret which is a byproduct of urea is structurally similar to biguanide. Plants have been the source of medicinal treatments for thousands of years. They play an essential role in the primary health care of 80% of the world’s developing and developed countries (King, Aubert, & Herman, 1998). Many of the currently available drugs have been directly or indirectly derived from plants. Out of an estimated 250,000 higher plants, less than 1% have been exploited pharmacologically and even less in regard to diabetes. There are about 800 plants that may possess potential antidiabetic ingredients. Functional plant foods such as fenugreek, curry leaves, bitter gourd, garlic, and sweet potato are few among 45 plants that have shown experimental or clinical antidiabetic activity (Grover

V. Perla, S.S. Jayanty / Food Chemistry 138 (2013) 1574–1580

1575

Fig. 1. Biguanides and related compounds. Guanidine is the basic unit of biguanide and has antidiabetic properties. It is structurally similar to urea. Biuret, a byproduct of urea, is structurally similar to biguanide. Galegine is an antidiabetic guanidine compound identified in French lilac known to have weight reducing effects. L-Arginine is a guanidine containing amino acid known to stimulate the release of insulin hormone. Metformin and phenformin are biguanides used as oral antihyperglycaemic drugs for type-2 diabetes treatment.

et al., 2002). Alkaloids, glycosides, galactomannan, polysaccharides, peptidoglycans, hypoglycans, guanidine, steroids, carbohydrates, glycopeptides, terpenoids, flavonoids, phenolics, amino acids and inorganic ions are a few of the plant-derived compounds that have demonstrated antidiabetic activity (Grover et al., 2002; Jung et al., 2006). The following active ingredients were found to be effective in controlling glucose levels in various diabetic experimental models (rats, rabbits, dogs, gerbils, langurs or humans): allicin and its sulphur containing amino acid precursor from garlic; polypeptide-p, momordicin, charantin (insulin like protein called plant insulin), kakara-1b, -111a and -111b, galactose-binding lectin and oleanolic acid 3-O-glucuronide from bitter gourd; and alkaloid trigonelline, furostanol saponins (trigoneosides glycoside D and trigofaenoside A), steroidal sapogenins (diosgenin and yamogenin), and 4-hydroxyisoleucine from fenugreek (Grover et al., 2002; Saxena & Vikram, 2004). Some flavonoids, polyphenols, and their sugar derivatives have been reported to be effective against the inhibitory activities of a-glucosidase and aldose reductase (Jung et al., 2006). Although, fenugreek, curry leaves, bitter gourd, garlic and sweet potato are known for antihyperglycaemic properties in Ayurvedic medicine for several centuries, detailed list of active components remain to be determined. The concept of synthesising biguanides, such as metformin, originated from the hypoglycaemic properties of plant guanidines. However, there has been no report on plant biguanides. The objective of this study was to quantify biguanide related compounds (BRCs) in functional plant foods (fenugreek, curry leaves, bitter gourd, garlic and sweet potato) and potato tubers. The presence of antidiabetic compounds, such as flavonoids and polyphenols, in potato tubers encouraged us to include them in this study (Perla, Holm, & Jayanty, 2012). 2. Materials and methods 2.1. Plant samples Dry fenugreek seeds (Trigonella foenum-graecum L.; Fabaceae family; 200 g), fresh green curry leaves (Murraya koenigii (L.) Sprengel; Rutaceae family; 170 g) and fresh green bitter gourds (Momordica charantia Descourt.; Cucurbitaceae family; 25 fruits) were purchased from Asian Indian Grocery store, Denver, Colorado. Garlic cloves (Allium sativum L.; Amaryllidaceae family; 16 bulbs)

and orange fleshed sweet potatoes (Ipomoea batatas (L.) Lam.; Convolvulaceae family; 6 tubers) were purchased from local markets. White-fleshed potatoes (Solanum tuberosum L. cv. Rio Grande Russet; Solanaceae family; 5 tubers) were collected from stored tubers harvested in September, 2010 at San Luis Valley Research Center. All the food samples, with the exception of fenugreek seeds, were freeze-dried at 0.12 mBar vacuum and 50 °C in a laboratory freeze-dryer (FreeZone 6, Labconco, Kansas City, MO). Freeze-dried samples and fenugreek dry seeds were ground to a fine powder in a coffee-grinder and stored separately in air-tight plastic zipper bags at 80 °C until further analysis. 2.2. Extraction of BRCs Each plant food was extracted three independent times using a method that was previously adopted for animal plasma with modifications (Freedman, Blitz, Gunsberg, & Zak, 1961). In each extraction, 15 ml of 20% NaCl was mixed with 1.67 g of powdered sample in a 250 ml capacity glass bottle and shaken horizontally at 250 rpm for 30 min at 25 °C. To this mixture, 3.75 ml of 50% trichloroacetic acid was added and mixed by vortex. After 10 min at room temperature, 190 ml of chloroform–methanol (85:15, v/ v) and 7.6 ml of 10 N NaOH were added and the mixture was shaken horizontally at 150 rpm for 25 min at 25 °C. The mixture was then centrifuged at 1960g for 10 min at room temperature. After aspirating and discarding the top aqueous phase and debris, the chloroform extract was collected in 1 l glass bottle. Three independent chloroform extractions from each food sample were pooled together in a 1 l bottle, reduced by evaporation to approximately 30 ml under a stream of nitrogen gas in a hood, and filtered with Whatman 40 filter paper. The concentrated chloroform extract was acidified with 0.4 ml of an acidic ethanol solution (2 N HCl in 95% ethanol) and evaporated to dryness under a stream of nitrogen gas in the hood. The dried extract was dissolved in 6 ml of a 5% NaCl solution by vortex and diluted with 6 ml of distilled water. After filtering, the final volume of the extract was adjusted to 12.5 ml with 2.5% NaCl solution and stored at 20 °C until further use. A zeolite base-exchange column was also used to isolate BRCs from fenugreek samples according to the procedure reported by Freedman et al. (1961) with modifications. Zeolite was purified according to the procedure of Freedman et al. (1961) and dried

1576

V. Perla, S.S. Jayanty / Food Chemistry 138 (2013) 1574–1580

at room temperature overnight. The base-exchange column was prepared by adding dry, purified zeolite powder (2 or 4 g) into a syringe barrel (2 cm diameter, 9 cm length, and 20 ml in volume) with filter paper circle (Whatman 40) as a sieve at the bottom of the barrel. The concentrated chloroform extract from 3 independent extractions of fenugreek samples was allowed to percolate through the zeolite base-exchange column by gravity. Then, the column was washed with 6 ml of 95% ethanol by gravity. BRCs were eluted with 6 ml of a 5% NaCl solution followed by 6 ml of distilled water by applying pressure with plunger. The column was not allowed to dry between adsorption, washing and elution steps. After the eluate was filtered, the final volume was adjusted to 12.5 ml with a 2.5% NaCl solution, and stored at 20 °C until further use. 2.3. Determination of BRCs BRCs in the plant food extracts were determined using the spectrophotometric method modified for 96-well microplates (Freedman et al., 1961; Heuclin, Pene, Savouret, & Assan, 1975). This assay is based upon the reaction of the guanidine group with anaphthol-diacetyl at an alkaline pH (Voges–Proskauer or V–P reaction). In brief, 200 ll of 3 N NaOH solution was mixed with 1 ml of sample extract, standard (1,1-dimethylbiguanide hydrochloride; also known as metformin HCl; in a 2.5% NaCl solution), or blank (2.5% NaCl solution) in a 2 ml microcentrifuge tube. The freshly prepared a-naphthol-diacetyl reagent (200 ll) was mixed with the contents of the tube and incubated in dark for 45 min at 37 °C. After incubation, approximately 5 mg of CeliteÒ S was mixed in each tube and centrifuged at 16.3g for 30 s at room temperature. Within 5 min after incubation, the supernatant from each tube was transferred to 3 wells (200 ll well1) in a 96-well flat bottom microplate (Costar 3370) and the absorbance was measured at 550 nm in a plate reader (Power Wave XS2, BioTek Instruments, Winooski, VT). Metformin HCl in 2.5% NaCl solution was used as the standard for calibration curves (Fig. 2B), and the total amounts of BRCs were expressed as metformin equivalents (ME) per g dry weight of the sample. Sample extracts were also tested by substituting the a-naphthol-diacetyl reagent with an equal amount of n-propanol to determine whether there is any interference of colour pigments of the sample in the assay (Sample blanks). Fresh anaphthol-diacetyl reagent was prepared by mixing 1.25 ml of 1% diacetyl solution in n-propanol with 10 ml of 25% a-naphthol in n-propanol, and the resulting solution was diluted to 50 ml with n-propanol.

Absorption spectra of metformin (36 lg ml1 in a 2.5% NaCl solution), curry leaf extract, and curry leaf extract spiked with metformin (1:1, v/v) were collected in the visible wave length (400– 700 nm) at 1 nm intervals using the plate reader (Fig. 2A). 2.4. Determination of L-arginine, urea and ammonia Urea as well as L-arginine and ammonia in the sample extracts were identified and quantified by a commercial kit (K-LARGE, Megazyme, Co., Wicklow, Ireland; Dukes & Butzke, 1998; Martin, Brandriss, Schneider, & Bakalinsky, 2003; Mira de Orduna, 2001). In principle, L-arginine was hydrolysed to urea and ornithine by the enzyme arginase. The urea produced was hydrolysed to ammonia (NH3) and carbon dioxide (CO2) by the enzyme urease. In the presence of glutamate dehydrogenase (GIDH) and reduced nicotinamide-adenine dinucleotide phosphate (NADPH), ammonia (in the form of ammonium ions, NHþ 4 ) reacts with 2-oxoglutarate to form + + L-glutamic acid and NADP . The amount of NADP formed was stoichoimetric with the amount of ammonia. For each mole of L-arginine or urea hydrolysed, two moles of NADPH were consumed. NADPH consumption was measured by the decrease in absorbance at 340 nm. The assay was performed after the pH of the sample extract or blank (2.5% NaCl solution) was adjusted to approximately 8.0 by 1 N NaOH solution. Then, 0.5 ml of sample extract or blank was mixed with 1.6 ml of distilled water and 0.5 ml of NADPH (in the buffer supplied by the vendor) in a disposable plastic cuvette (3 ml capacity). After approximately 2 min, absorbance of the solution (A1) was measured at 340 nm using a spectrophotometer (DU 720, Beckman-Coulter, Fullerton, CA). Reaction was initiated immediately by mixing of 0.02 ml of GIDH, and the absorbance of the solution (A2) was measured after approximately 2 min. Then, 0.05 ml of urease was mixed and the absorbance of the solution (A3) was measured after approximately 6 min. Finally, 0.02 ml of arginase was mixed and the absorbance of the solution (A4) was measured at the end of reaction after approximately 7 min. The absorbance difference (A1  A2) for both blank and sample extract was determined, and DAammonia values were obtained by subtracting absorbance difference of blank from the absorbance difference of the sample extract. Similarly, absorbance differences [(A2  A3) and (A3  A4)] for both blank and sample extracts were determined, and DAurea and DAL-arginine values were estimated, respectively. The concentrations of ammonia, urea and L-arginine in the sample extracts were calculated using the following formula, and the final values were expressed in lg ml1:

Fig. 2. (A) Absorption spectra of biguanides in the visible wave length; (B) Standard curve of metformin HCl. Each value in both the figures is a mean of six determinations.

V. Perla, S.S. Jayanty / Food Chemistry 138 (2013) 1574–1580 1

Concentration ðg l Þ ¼ ðFinal volume; mlÞ 1

 ðMolecular weight; g mol Þ

1577

cal analyses were performed using XLSTAT-Pro software, version 2011.4.04 (Addinsoft USA, New York, NY).

 ðDAÞ=edv where, e is 6300 (the extinction coefficient of NADPH at 340 nm; l mol1 cm1); d is the length of light path (1 cm); and v is the sample volume (ml). Molecular weights of ammonia, urea and L-arginine used were 17.03, 60.06 and 174.21, respectively. 2.5. Estimation of corrections for BRCs in the sample extracts Based on the concentration range of L-arginine, urea and ammonia in the sample extracts, calibration curves for these compounds were developed in a 2.5% NaCl solution using the V–P assay as described above (Figs. 3 and 4). Calibration curves were prepared only when the compounds tested responded in the V–P reaction of the biguanide assay. From the calibration curves, absorbance values for L-arginine, urea and ammonia in the sample extracts were obtained and the percentage of interference due to the presence of L-arginine, urea or ammonia in the biguanide assay was determined using the following formula:

Interference ð%Þ ¼ ðAinterfering compound =Atotal BRCs Þ  100 where, A is the absorbance at 550 nm. We observed that ammonia and spermidine (a polyamine structurally similar to biguanides) in a 2.5% NaCl solution at different concentrations did not interfere with the V–P reaction (Data not shown). For this reason, we did not pursue these compounds further. Total interference in the V–P assay was estimated by adding percentages of interference due to L-arginine and urea in the sample extracts. Corrected values of BRCs were obtained after subtracting the total interference values from the total values of BRCs in each sample. 2.6. Statistical analysis Each replication in an assay represents 3 independent extractions. Unless otherwise mentioned, each assay was performed twice with 3 replications. Six observations (n = 6) for each assay were statistically analysed. The results were expressed as the mean ± standard deviation (SD). Significant differences among the means were determined using one-way analysis of variance (ANOVA) followed by Tukey’s Honestly Significant Differences (HSD) multiple-rank-test at p 6 0.05 significance level. All statisti-

3. Results and discussion 3.1. Absorption spectra of BRCs Absorption spectra of metformin, curry leaf extract, and curry leaf extract spiked with metformin in the visible region in the V– P assay are presented in Fig. 2A. Metformin exhibited maximum absorption between 546 to 558 nm in a 2.5% NaCl solution. Based on this range, the wave length of 550 nm was considered to be the lambda max (kmax) for metformin. Similarly, metformin spiked with equal volume of curry leaf extract, or curry leaf extract alone exhibited kmax at approximately 550 nm. Based on these results, a calibration curve for metformin was prepared at 550 nm (Fig. 2B). Previous studies on phenformin recovery from biological samples, such as plasma, liver, kidney, muscle tissues and urine, employed V–P assays at 565 nm (Freedman et al., 1961; Heuclin et al., 1975). The kmax employed for metformin in the present study was close to 565 nm, and the deviation in kmax was attributed to the composition of biguanide tested and plant species.

3.2. L-Arginine and urea in the sample extracts The extraction of biguanides from biological samples with chloroform–methanol solvent eliminates common interfering compounds, such as L-arginine and urea, in the final sample extract (Heuclin et al., 1975). However, in the present study, trace amounts of L-arginine were present as evidenced by the enzymatic assays of the sample extracts, ranging between 0.06 (sweet potato) to 0.41 lg ml1 (curry leaves) (Fig. 3A). With the exception of sweet potato, L-arginine levels were not significantly different from each other in the sample extracts. Because L-arginine is a guanidino amino acid, it is considered to be among the BRCs. L-Arginine is also one of the major interfering compounds in V–P assay (Barritt, 1936; Heuclin et al., 1975). For these reasons, L-arginine was quantified separately by an enzymatic assay and its interference was eliminated during the quantification of BRCs by the V–P assay. These enzymatic assays were also determined the lowest and highest amounts of urea in the sample extracts of sweet potato (0.06 lg ml1) and curry leaves (0.31 lg ml1), respectively (Fig. 4A). All other extracts contained about 0.1 to 0.2 lg ml1 of urea and were not significantly different from each other.

Fig. 3. Interference of L-arginine in the V–P assay. (A) Amount of L-arginine in the sample extraction. (B) Standard curve of L-arginine obtained with the V–P assay. (C) Percentage of interference due to L-arginine in the V–P assay of plant foods. Bars are means ± SD of four determinations. The mean values that are not significantly different are represented by same letter (p > 0.05).

1578

V. Perla, S.S. Jayanty / Food Chemistry 138 (2013) 1574–1580

Fig. 4. Interference of urea in the V–P assay. (A) Amount of urea in the sample extraction. (B) Standard curve of urea obtained with the V–P assay. (C) Percentage of interference due to urea in the V–P assay of plant foods. Bars are means ± SD of four determinations. The mean values that are not significantly different are represented by same letter (p > 0.05).

3.3. Corrections for BRCs in the plant foods Because chloroform–methanol extraction did not completely eliminate L-arginine and urea, corrections for these common compounds were estimated in each plant food and subtracted from their respective total BRCs in each plant food. The response of Larginine in the V–P reaction of the biguanide assay was positive and exhibited linear relationship between the absorbance values and the concentrations tested (0, 0.2, 0.4 and 0.6 lg ml1) (Fig. 3B). This concentration range falls within the range of L-arginine in the sample extracts (0.06–0.41 lg ml1; Fig. 3A). This data suggests that traces of L-arginine present in the sample extracts interfered with the V–P assay. This interference was very high in the garlic and sweet potato sample extracts (60–70%) compared to that of other sample extracts (14–27%) (Fig. 3C). Similarly, the response of urea in the V–P reaction was positive and linear within the concentration range of urea in the sample extracts (Fig. 4A and B). The highest percentage of interference in the V–P assays due to urea was measured for sweet potato (46%) followed by garlic (21%) (Fig. 4C). All of the other samples exhibited 4–9% of urea interference in the V–P assays and were not significantly different from each other. Urea is not a guanidine compound. But, it interfered in the V–P reaction due to structural similarity with guanidine (Barritt, 1936) (Fig. 1). 3.4. BRCs in the plant foods The BRCs in the plant foods before and after correcting for common interfering compounds (L-arginine and urea) in V–P assay are

presented in Fig. 5A. The trend remained same after the BRCs of all the plant species tested were corrected. Sweet potato and garlic did not contain significant amount of BRCs after corrections (60.4 lg g1). On the other hand, the highest amount of BRCs were present in curry leaves (25.57 lg g1) followed by fenugreek seeds (18.98 lg g1) and bitter gourd (11.14 lg g1). Presence of BRCs and other antidiabetic compounds discussed earlier, support experimental or clinical anti-diabetic activity of these functional foods (Grover et al., 2002). Interestingly, potato tubers, a carbohydrate-rich plant food, also contained significant amount of BRCs (7.05 lg g1). Because the V–P reaction is specific to guanidine group, these plant foods most likely contained guanidine, biguanide and related compounds after corrections. Although biguanides in plants have not been reported, many kinds of guanidine compounds, including guanidinosuccinic acid, creatine, guanidinoacetic acid, Na-acetyl-L-arginine, b-guanidinopropionic acid, cguanidinobutyric acid, L-arginine, c-guanidinobutyramide, guanidine, creatinine, taurocyamine, and methylguanidine, were identified in several other plant species (Kato, Kondo, & Mizuno, 1986) (Fig. 6). Guanidine and biguanides are known for their antidiabetic properties (Bailey & Day, 2004). However, we are not sure whether all the derivatives of guanidine possess similar antidiabetic properties. Notably, some of the guanidine derivatives, such as methylguanidine, are toxic in nature (De Jonge et al., 2001). The enzymes urease and glutamate dehydrogenase used in the KLARGE assay are specific for their substrates urea and ammonia, respectively (Gutmann & Bergmeyer, 1974). But, there are several substrates for arginase. If these substrates are present in the sample extract, an arginase-based enzymatic assay may yield false-po-

Fig. 5. (A) Biguanide and related compounds (BRCs) in the plant foods. (B) Effect of extraction method on the yield of BRCs from fenugreek seeds. Bars are means ± SD of six determinations. The mean values that are not significantly different are represented by same letter (p > 0.05). ME: metformin equivalents; DW: dry weight.

V. Perla, S.S. Jayanty / Food Chemistry 138 (2013) 1574–1580

1579

while using the V–P assay and the correction procedures in plant species, especially when they contain very low amounts of BRCs. 3.5. Effect of zeolite column on the extraction of BRCs A base-exchange column containing 4 g of zeolite did not extract significantly higher amounts of BRCs from fenugreek seeds compared to using 2 g of zeolite (Fig. 5B). However, acid ethanol extracted approximately 5-fold more BRCs (23.52 lg g1) from fenugreek seeds compared to that of zeolite columns (4.47– 5.05 lg g1). Freedman et al. (1961) found that a Folin Decalso (zeolite) adsorption column gave higher yields when it was used to isolate phenformin from animal tissues spiked with it. In addition, phenformin extraction with acid ethanol method also gave good results with plasma spiked with phenformin (Freedman et al., 1961; Heuclin et al., 1975). However, in the present study, the results from the V–P assays suggest that more BRCs from plant foods are extracted with acid ethanol method than zeolite column. 3.6. Pathways involved in biguanide synthesis in the plant foods Fig. 6. Possible metabolic pathways involved in biosynthesis of biguanide in the plant foods. Established metabolic pathways in animals or plants are indicated by solid arrows. Proposed route(s) of biguanide biosynthesis in plant foods are indicated by broken arrows. GABA: c-aminobutyric acid; d-AVA: d-aminovaleric acid; b-GPA: b-guanidinopropionic acid; c-GBA: c-guanidinobutyric acid; d-GVA: dguanidinovaleric acid; AGAT: L-arginine-glycine-transamidinase; GMT: guanidine methyltransferase; TA: transamidination. Refer text for details.

sitive results for L-arginine (Mira de Orduna, 2001). Hepatic arginase of ureotelic and ureocotelic animals have nearly same substrate specificity toward L-arginine and to some extent Lcanavanine (Mora, Tarrab, Martuscelli, & Soberön, 1965). Leupine arginase hydrolyses L-homoarginine more efficiently than L-arginine (Muszynska & Reifer, 1968). In addition to L-arginine, significant amount of agmatine and L-canavanine were reported to hydrolyse by arginase from Vigna catjang. Moreover, buffalo liver arginase hydrolysed only L-arginine and L-canavanine. Interestingly, the arginase enzyme did not hydrolyse D-arginine (Dabir, Dabir, & Somvanshi, 2005). Proteins, arginine, agmatine, creatine, creatinine, guanidine acetic acid, guanidine HCl, methylguanidine, aminoguanidine, urea and biuret produce positive results in the V–P test. Whereas, amino acids (except arginine), uric acid, allantoin, hydantoin, adenine, xanthine, guanine, methyl urea, common monosaccharides, disaccharides, polysaccharides, glucosides and various organic acids yielded negative results in the V–P test (Barritt, 1936). Heuclin et al. (1975) found that creatinine, uric acid, urea, and arginine did not cross-react with the a-naphthol-diacetyl reagent in the V–P assay when samples were extracted with chloroform–methanol. In the present study, trace amounts of arginine and urea found after chloroform–methanol extraction in the sample extracts were eliminated by proper correction procedures adopted during V–P assay. Proteins in the food samples were also eliminated due to the presence of large quantities of chloroform–methanol and trichloroacetic acid in the solvent during the extraction of BRCs (Fic, Kedracka-Krok, Jankowska, Pirog, & Dziedzicka-Wasylewska, 2010). The alternative substrates of arginase that interfere with the V–P assay were also eliminated along with L-arginine after correction. In the V–P assay, intensity of the pink colour and corresponding absorbance values depend upon the type of molecule present and its concentration in the assay mixture. This is evident from the calibration curves of L-arginine and urea (Figs. 3B and 4B). L-Arginine and urea in the same concentration range yielded different absorbance values in V–P assays. Thus, caution should be taken

De Jonge et al. (2001) investigated the accumulation of guanidine compounds in mice (arginine–guanidine pathway). Guanidine compounds are the metabolites of arginine that retain the guanidinium group. In arginine–guanidine pathway, the guanidino group of arginine is transamidinated to glycine to yield guanidinoacetic acid and subsequently creatine and creatinine. After reacting with an oxygen radical, creatinine is converted to methylguanidine, which is further modified to guanidine (De Jonge et al., 2001) (Fig. 6). Free guanidines are widespread in plants (Reinbothe & Mothes, 1962). In the arginine–guanidine pathway, in addition to the physiological substrate glycine, its homologues, such as b-alanine, c-aminobutyric acid and d-aminovaleric acid, can serve as substrates for L-arginine-glycine transamidinase, yielding b-guanidinopropionic acid, c-guanidinobutyric acid and d-guanidinovaleric acid, respectively (De Jonge et al., 2001). In several other plant species, studies were conducted to understand metabolic routes of arginine other than the urea cycle. Kato et al. (1986) reported the existence of an arginine metabolic pathway via c-guanidinobutyramide and c-guanidinobutyric acid; and a metabolic pathway of an arginine via guanidinoacetic acid and creatine to creatinine in certain plant species or during certain stages of life cycle in some plant species. Biochemically, guanidine can serve as an immediate precursor for biguanide biosynthesis. The identification of BRCs and existence of the arginine–guanidine pathway in the plant species support this proposed route of biguanide biosynthesis in the plant foods studied (Fig. 6). Galegine, a guanidine derivative, has been shown to be synthesised in the seedlings, leaves, flowers, and fruits of Galega officinalis (Reuter, 1962). Arginine can transfer its amidine group to a precursor of galegine by a transamidination reaction (Fig. 6). A transamidination reaction could possibly occur between arginine and isopentenylamine (Reuter, 1962). Similarly, galegine may serve as immediate precursor for biguanide biosynthesis in the plant foods. In this study, we determined that all of the plant foods tested contained urea. Urea is formed during the enzymatic degradation of arginine and purines. Arginine is converted into ornithine and urea by the enzyme arginase. Purines are degraded via allantoin and allantoic acid, and produce glyoxylic acid and urea in plants (Reinbothe & Mothes, 1962; Thomas & Schrader, 1981, Fig. 6). Arginases from some plant species can also utilise substrates, such as agmatine and L-canavanine, to produce urea (Dabir et al., 2005). Furthermore, toxic ammonia is detoxified to urea via urea–ornithine cycle. The urea produced in this cycle is speculated to be

1580

V. Perla, S.S. Jayanty / Food Chemistry 138 (2013) 1574–1580

mainly associated with amino acid synthesis and is not considered to be the main source of urea production (Reinbothe & Mothes, 1962). There are more than one possible routes of urea production in plants (Fig. 6). Accumulated urea in plants is utilised in two ways. In one group of plants, urea is hydrolysed by the enzyme urease. These plants are widespread in the plant kingdom. Nonurease plants use urea in other ways (Reinbothe & Mothes, 1962). Whether a plant is urease or non-urease type, the presence of urea and BRCs in all the plant foods tested suggest that some percentage of urea might be converted to biguanide. Although, biuret is structurally similar to biguanide, we are not sure whether biuret is involved in this pathway. Biuret may serve as an intermediate compound during the biosynthesis of biguanide from urea in these plant foods (Fig. 6). Enzymatic conversion of biuret into other compounds was well documented in bacteria (Cameron, Durchschein, Richman, Sadowsky, & Wackett, 2011; Cook, Beilstein, Grossenbacher, & Hütter, 1985). Similarly, enzymatic conversion of biuret into other intermediate metabolites of the biguanide pathway can’t be ruled out in the plant species. One or more of the proposed pathways of biguanide biosynthesis are most likely operating in the plant foods tested in this study. These proposed pathways are further supported by the following facts: (1) chemically, the structure of biguanide is considered to be a condensed system of amidino-groups or of guanidine-units; (2) biguanides are indirectly related to guanidine and urea because of their close relationship with amidino(thio)ureas, cyanoguanidines and biurets; (3) this relationship is evident from the existence of comparable synthesis, and general physical and chemical behaviour of the relevant compounds; and (4) chemical synthesis of biguanides are usually performed by using urea compounds, such as O-alkylisoureas and O-methyl-amidinoisourea (Kurzer & Pitchfork, 1968). In addition, sulphonylureas are also the urea derivatives that are commonly used for the treatment of type-2 diabetes (Mahajan & Gupta, 2010). Based on the results of this study, three routes of biguanide biosynthesis are proposed. In addition to several possible intermediate metabolites, guanidine, urea and/or galegine may play key roles in these proposed biguanide biosynthetic pathways in the plant foods. 3.7. Conclusion BRCs were quantified in fenugreek seeds, green curry leaves, green bitter gourd, garlic, sweet potato and potato tubers. The highest amount of BRCs was found in curry leaves followed by fenugreek, bitter gourd and potato. Garlic and sweet potato contained very low or negligible amounts of BRCs. If proper corrections for interfering compounds are employed, the V–P assay can be successfully used to quantify BRCs in plant species. We propose that biguanide is synthesised from guanidine, urea and/ or galegine in the plant foods. Further investigations are required to confirm these proposed biosynthetic routes and identify intermediate metabolites, enzymes and genes involved in these pathways. Acknowledgments We thank Dr. Henry J. Thompson and Dr. David G. Holm for their suggestions. This work was partially supported by a grant from the Colorado Department of Agriculture through the USDA’s Specialty Crop Block Grant Program (award #10991) and Colorado Potato Administrative Committee Area II.

References Bailey, C. J., & Day, C. (2004). Metformin: Its botanical background. Practical Diabetes International, 21, 115–117. Barritt, M. M. (1936). The intensification of the Voges–Proskauer reaction by the addition of a-naphthol. The Journal of Pathology and Bacteriology, 42, 441–454. Cameron, S. M., Durchschein, K., Richman, J. E., Sadowsky, M. J., & Wackett, L. P. (2011). New family of biuret hydrolases involved in s-Triazine ring metabolism. ACS Catalysis, 1, 1075–1082. Cook, A. M., Beilstein, P., Grossenbacher, H., & Hütter, R. (1985). Ring cleavage and degradative pathway of cyanuric acid in bacteria. Biochemical Journal, 231, 25–30. Dabir, S., Dabir, P., & Somvanshi, B. (2005). Purification, properties and alternate substrate specificities of arginase from two different sources: Vigna catjang cotyledon and buffalo liver. International Journal of Biological Sciences, 1, 114–122. De Jonge, W. J., Marescau, B., D’Hooge, R., De Deyn, P. P., Hallemeesch, M. M., Deutz, N. E. P., et al. (2001). Overexpression of arginase alters circulating and tissue amino acids and guanidino compounds and affects neuromotor behavior in mice. The Journal of Nutrition, 131, 2732–2740. Dukes, B., & Butzke, C. (1998). Rapid determination of primary amino acids in grape juice using an o-phthaldialdehyde/N-acetyl-L-cysteine spectrophotometric assay. American Journal of Enology and Viticulture, 49, 125–134. Fic, E., Kedracka-Krok, S., Jankowska, U., Pirog, A., & Dziedzicka-Wasylewska, M. (2010). Comparison of protein precipitation methods for various rat brain structures prior to proteomic analysis. Electrophoresis, 31, 3573–3579. Freedman, L., Blitz, M., Gunsberg, E., & Zak, S. (1961). Determination of phenformin in biologic fluids and tissues. Journal of Laboratory and Clinical Medicine, 58, 662–666. Goo, A. K. Y., Carson, D. S., & Bjelajac, A. (1996). Metformin: A new treatment option for non-insulin-dependent diabetes mellitus. Journal of Family Practice, 42, 612–618. Grover, J. K., Yadav, S., & Vats, V. (2002). Medicinal plants of India with anti-diabetic potential. Journal of Ethanopharmacology, 81, 81–100. Gutmann, I., & Bergmeyer, H. U. (1974). Urea: Determination of urea with glutamate dehydrogenase as indicator enzyme. In H. U. Bergmeyer (Ed.), Methods of enzymatic analysis. Weinheim, Germany: Verlag Chemie. Heuclin, C., Pene, F., Savouret, J. F., & Assan, R. (1975). Characterization of phenformin and metabolites in plasma. Diabete & Metabolisme (Paris), 1, 235–240. Jung, M., Park, M., Lee, H. C., Kang, Y., Kang, E. S., & Kim, S. K. (2006). Antidiabetic agents from medicinal plants. Current Medicinal Chemistry, 13, 1203–1218. Kato, T., Kondo, T., & Mizuno, K. (1986). Occurrence of guanidine compounds in several plants. Soil Science & Plant Nutrition, 32, 487–491. King, H., Aubert, R. E., & Herman, W. H. (1998). Global burden of diabetes, 1995– 2025: Prevalence, numerical estimates, and projections. Diabetes Care, 21, 1414–1431. Kurzer, F., & Pitchfork, E. D. (1968). The chemistry of biguanides. Topics in Current Chemistry, 10(3), 375–472. Mahajan, R., & Gupta, G. (2010). Drugs in pipeline for type-2 diabetes. Internet Journal of Pharmacology, 8, 3-3. Martin, O., Brandriss, M. C., Schneider, G., & Bakalinsky, A. T. (2003). Improved anaerobic use of arginine by Saccharomyces cerevisiae. Applied and Environmental Microbiology, 69, 1623–1628. Mira de Orduna, R. (2001). Quantitative determination of L-arginine by enzymatic end-point analysis. Journal of Agricultural and Food Chemistry, 49, 549–552. Mooney, M. H., Fogarty, S., Stevenson, C., Gallagher, A. M., Palit, P., Hawley, S. A., et al. (2008). Mechanisms underlying the metabolic actions of galegine that contribute to weight loss in mice. British Journal of Pharmacology, 153, 1669–1677. Mora, J., Tarrab, R., Martuscelli, J., & Soberön, G. (1965). Characteristics of arginases from ureotelic and non-ureotelic animals. Biochemical Journal, 96, 588–594. Muszynska, G., & Reifer, I. (1968). Purification, properties and inhibition of plant arginase. Acta Biochimica Polonica, 15, 55–66. Perla, V., Holm, D. G., & Jayanty, S. S. (2012). Effects of cooking methods on polyphenols, pigments and antioxidant activity in potato tubers. LWT-Food Science & Technology, 45, 161–171. Reinbothe, H., & Mothes, K. (1962). Urea, ureides, and guanidines in plants. Annual Review of Plant Physiology, 13, 129–149. Reuter, G. (1962). On guanidine metabolism in Galega officinalis. Phytochemistry, 1, 63–65. Saxena, A., & Vikram, N. K. (2004). Role of selected Indian plants in management of type 2 diabetes: A review. Journal of Alternative and Complementary Medicine, 10, 369–378. Thomas, R. J., & Schrader, L. E. (1981). Ureide metabolism in higher plants. Phytochemistry, 20, 361–371. Tousoulis, D., Antoniades, C., Tentolouris, C., Goumas, G., Stefanadis, C., & Toutouzas, P. (2002). L-Arginine in cardiovascular disease: Dream or reality? Vascular Medicine, 7, 203–211.