- Email: [email protected]
Effect of neomycin on the hydrolysis and toxicity of vicine and convicine in rats
Fd Chem. Toxic. Vol. 31, No. I1, pp. 835-840, 1993
0278-6915/93$6.00+ 0.00 Copyright © 1993PergamonPress Ltd
Printed in Great Britain.All rights reserved
EFFECT OF NEOMYCIN ON THE HYDROLYSIS AND TOXICITY OF VICINE AND CONVICINE IN RATS M. S. S. ARBID,M. S. MADHYASTHA,R. R. MARQUARDT*and A. A. FROHLICH Department of Animal Science, The University of Manitoba, Winnipeg, Canada R3T 2N2 (Accepted 11 June 1993)
Abstract--This study in the rat established the effects that a broad-spectrum and poorly absorbed antibiotic, neomycin sulfate, had on the in vitro and in vivo hydrolysis of vicine and convicine by the intestinal microflora, and on vicine- and convicine-induced depletion of blood glutathione and the resulting toxicity. The in vitro studies demonstrated that digesta from the caecum and large intestine were highly effective in hydrolysing vicine and convicine, whereas digesta from the same sections of the gastro-intestinal tract of neomycin-treated rats were much less effective (P < 0.0001). The in vivo studies showed that the total amount of vicine and convicine excreted in the urine and faeces was much greater in neomycin-treated rats compared with controls (P < 0.05), indicating the ability of neomycin to increase the amount of glycosides, particularly that of vicine, excreted in the faeces. The ability of glycosides to decrease the concentration of glutathione in blood (P < 0.05) and to increase rat mortality was greatly reduced in rats that were treated with neomycin, particularly in those treated ip with the toxin. Thus, the results demonstrated that neomycin reduced the rate at which vicine and convicine were hydrolysed in the lower section of the gastro-intestinal tract, and that neomycin treatment was associated with a reduced toxicity of the glycosides.
INTRODUCTION Vicine and convicine are glycosides that are found primarily in the faba bean (Vicia f a b a L.), which is one of the most important pulse crops in the world (Marquardt, 1989). Vicine and convicine are hydrolysed by the intestinal microflora (Frohlich and Marquardt, 1983; Hegazy and Marquardt, 1984) to the highly reactive free-radical-generating compounds (Albano et al., 1984) divicine and iso-uramil. Divicine and iso-uramil have been strongly suggested to be the causative agents in favism (Belsey, 1973; Bottini, 1973; Donoso et al., 1969), a haemolytic disease in humans, particularly in young males, that have a deficiency in erythrocytic glucose-6-phosphate dehydrogenase activity (Beutler, 1978; Yoshida, 1973). These free-radical generators may also cause other adverse effects including lipid peroxidation (D'Aquino et al., 1983), altered fat (Muduuli et al., 1982) mitochondrial metabolism (Graf et al., 1985), glutathione oxidation (Mager et al., 1969) and possibly diabetes (Rocic et al., 1985). Frohlich and Marquardt (1983) showed that a small percentage of dietary vicine, but not convicine was absorbed from the gastro-intestinal tract and excreted through the bile and kidney in chickens. Further, they demonstrated that the glycosides were not hydrolysed in vitro by avian tissue homogenates but were rapidly hydrolysed by digesta from the caecum of the birds. The total faecal excretion of
*To whom correspondence should be addressed. Abbreviation: GSH = glutathione.
vicine and convicine was increased when antibiotics were added to the diet. Similar results were obtained with the rat (Hegazy and Marquardt, 1984). In other studies, Arbid and M arquardt (1986) and Yannai and Marquardt (1985) demonstrated in rats that the route of administration of vicine and convicine greatly affected their toxicities. Oral administration of relatively large amounts of the glycosides to rats produced only small reductions in blood glutathione (GSH) concentrations and did not cause mortality. In contrast, ip administration of these compounds to rats resulted in a rapid decrease in the concentration of GSH in blood followed by death caused by anoxia. Rats that died from the toxic effects of these compounds had a large proportion of erythrocytes that were not able to exchange respiratory gases (Arbid and Marquardt, 1986 and 1988). These results cannot, however, be attributed to the direct hydrolysis of vicine and convicine by animal tissues (Frohlich and Marquardt, 1983; Hegazy and Marquardt, 1984) or to a direct effect of vicine and convicine on erythrocytes (Arbid and Marquardt, 1986). The results of the above studies indicated that, compared with the oral route, ip administration of vicine and convicine to rats resulted in a rapid uptake and hydrolysis of these compounds in the gastrointestinal tract, followed by the absorption of the aglycone into blood in sufficient quantities to overwhelm the protective effects of the GSH reductase system. Data from the literature therefore provide indirect (and not direct) evidence that vicine and convicine must first be hydrolysed to their aglycones to become toxic.
835
836
M.S.S. ARBIOet al.
The current study was based on the hypothesis that the toxicity of vicine and convicine administered ip to rats that are treated orally with neomycin will be decreased since the antibiotic will inhibit the growth of anaerobic micro-organisms in the gastro-intestinal tract, and consequently the decreased production of the aglycones divicine and iso-uramil will in turn result in a less marked effect on blood GSH concentration and lower mortality. The purpose of this study was to provide more definitive evidence to support this hypothesis. In order to achieve this goal, vicine and convicine were administered by both the ip and oral routes to control and neomycin-treated rats, and the mortality, blood GSH concentration and degree of hydrolysis of the glycosides by digesta from the gastro-intestinal tract were monitored. MATERIALS AND METHODS
Animals and treatment
Sprague-Dawley male albino rats (150_+5g) used in in vivo studies were obtained from the animal house of the National Research Centre (Dokki, Cairo, Egypt). Sprague-Dawley male albino rats (180-250 g) used in in vitro studies were obtained from the breeding farm of the University of Manitoba (Winnipeg, MB, Canada). The rats were housed individually in metabolic cages (Maryland Plastics, Inc., Federalsburg, MD, USA) and were maintained on a standard diet and given water ad lib. Chemicals
Vicine and convicine were prepared as described by Marquardt et al. (1983). Neomycin sulfate tablets (500 g each) were from Memphis Co. (Cairo, Egypt). Neomycin sulfate (powder) and fl-glucosidase were obtained from Sigma Co. (St Louis, MO, USA). Other chemicals were from Fisher Scientific Ltd (Winnipeg, MB, Canada). Estimation of vicine and convicine in urine, faeces and digesta
1 ml suspension of digesta or faeces in 0.145M saline (1 g/10 ml) was mixed with 5 ml perchloric acid (6%) and vortexed for 1 rain. Samples were passed through 0.5 ~m-filter for the clean-up. Vicine and convicine concentrations in these samples were measured by HPLC using a modification of the procedure described by Marquardt and Frohlich (1981). A Waters 600E Multisolvent Delivery HPLC System (Canada Waters Chromatography Division, ONT, Canada) with a 20-/~1 loop injector, a reversephase column (type Ultremex 5C18, 25 cm x 4.6 mm i.d., 5/~m; Phenomenex, Rancho Palos Verdes, CA, USA) and a variable-wavelength UV detector (Waters 484; Waters Associates Milford, MA, USA) set to 282nm and 0.01 units of sensitivity (absorbance units full scale or AUFS) was used. Glassdistilled water was used in the isocratic mobile phase in place of the eluting solution used by Marquardt
and Frohlich (1981). Divicine and iso-uramil were monitored in the absence of oxygen at 280 nm at a low pH (Marquardt and Frohlich, 1981). In vitro studies Hydrolysis of vicine and convicine in digesta. The first experiment used three rats weighing about 250 g. The animals were killed and digesta from the stomach, small intestine, caecum and large intestine were removed from each rat separately and weighed. Samples of digesta (! g each) from different parts of the gastro-intestinal tract were transferred separately and in duplicate from each of the three rats into capped test-tubes containing 10 ml 0.145 M degassed saline. The mixture was vortexed for i min, after which half of the volume in each tube was withdrawn and an equal volume of 0.145M degassed saline containing vicine (!.5 mg) and convicine (0.75 mg) was added. The mixture was incubated at 37°C for 10, 20, 40 and 60 min in a shaking water-bath. In the above studies, the saline solution was degassed by bubbling with oxygen-free nitrogen for 10 min. Incubation of digesta with vicine and convicine resulted in the maintenance of anaerobic conditions for at least 60min, and probably longer, as indicated by the persistence of the hydrolytic compounds divicine and iso-uramil, which have been shown to be readily inactivated by the presence of a trace amount of dissolved oxygen (Hegazy and Marquardt, 1984; Marquardt, 1989). Similar anaerobic in vitro fermentation procedures have been used to monitor the hydrolysis of dietary fibre by ileal digesta from pigs (McBurney and Sauer, 1993). In a second experiment, six rats weighing about 180-240 g were divided into two groups of three each and were housed individually in metabolic cages. The first group (control) was fed a normal stock diet, and the second group was given the diet containing 5% neomycin. The rats were killed after 3 days of feeding, and digesta from the caecum and large intestine of each rat (1 g each) were incubated with vicine and convicine for 30min as described above. In vivo studies Effect of neomycin on the hydrolysis of vicine and convicine. In the third experiment 10 male albino Sprague-Dawley rats weighing 150 + 5 g were divided into two groups of five each and were fed the standard diet containing 1% vicine and 1% convicine. The first group served as the control and the second group was treated by intubation with neomycin (250 mg/kg body weight, four times a day) for 24 hr prior to and during the collection period. Urine and faecal samples were collected over the following 48-hr period, and the pooled samples were assayed for vicine and convicine. Effect of neomycin on the toxicity of vicine and convicine administered orally or ip to rats. In the fourth experiment, 50 rats weighing 150 _+ 5 g were divided into 10 groups of five rats each. The 10
Vicine and convicine hydrolysis in rats B
A
RESULTS
C
HPLC assay for vicine and convicine
3
OJ
3
2
"6 O) ¢.) C
!
I
I
I
I
837
I
I
0 I 23456
!
I
!
I
0 12345
I
I
I
I
I
I
I
I
I
I
601254567
Retention time (min)
Fig. 1. Chromatograms of vicine and convicine incubation mixtures. The hydrolysis and elution patterns of vicine and convicinewere obtained when the two toxins were incubated with digesta from the caecum for different periods: 0.0 min (A), 30 rain (B) and 60 min (C). The different peaks correspond to the solvent (I), convicine (2) and vicine (3). experimental treatments were as follows: (1) control, physiological saline administered orally; (2) control, physiological saline administered ip; (3) control, vicine administered orally; (4) control, vicine administered ip; (5) control, convicine administered orally; (6) control, convicine administered ip; (7) neomycin administered orally followed by vicine given orally; (8) neomycin administered orally followed by vicine administered ip; (9) neomycin administered orally followed by convicine given orally; and (10) neomycin administered orally followed by convicine administered ip. Neomycin was administered to rats in groups 7 to 10 for 24 hr prior to administration of the glycosides and then throughout the 48-hr experimental period (250mg neomycin/kg body weight, four times a day). In all cases administration of the drugs was by intubation or ip injection. Vicine (5 g/kg body weight) and convicine (4g/kg body weight) were administered separately as indicated above. Rats that died in the groups treated ip with vicine and convicine (groups 4, 6, 8 and 10) were replaced with rats from the same stock colony. The same treatments were repeated with a sufficient number of rats so that the final total number of rats that survived for 48 hr was five per group. Blood samples from the orbital sinus of rats were collected in heparinized sterile tubes on ice, as described by Helperin et al. (1951), 2 hr after administration of vicine and convicine. Glutathione in blood was determined according to the procedure of Beutler (1963). Urine and faecal samples were collected for 24 hr after administration of the glycosides and were assayed for vicine and convicine as described above. Mortality represented the percentage of rats that died within 24 hr after vicine and convicine administration. Statistical analysis Data were subjected to analysis of variance (ANOVA), using the General Linear Model (GLM) procedure, and to regression analysis (Statistical Analysis System Institute Inc., 1986).
The assay procedure, which is a modification of that reported by Marquardt and Frohlich (1981), involved the elution of vicine and convicine in water instead of 0.05 M phosphoric acid-ammonia buffer (pH 2.0). The modified procedure improved the resolution of the two compounds, simplified the column clean-up and yielded a linear relationship between the peak areas of both vicine and convicine and the concentrations of the compounds used (i.e. up to 55/~g/ml for vicine and 25 #g/ml for convicine). A typical chromatogram for the hydrolysis of vicine and convicine is shown in Fig. 1.
In vitro hydrolysis of vicine and convicine The results of the first experiment demonstrated that the caecum and large intestine were the main sites for the hydrolysis of the two glycosides. The percentages of hydrolysis (_+ SE) of vicine and convicine, respectively, by digesta from different sections of the gastro-intestinal tract during a 60-min incubation period (at 37°C) were: 13.8 _+2.1 and 6.2 _+ 1.8 for the stomach, 27.8 _+ 1.2 and 45.4_+ 0.3 for the small intestine, 65.7 _+23.8 and 53.8 +_ 3.2 for the large intestine, and 77.7 + 0.1 and 85.25 _+0.25 for the caecum. For both glycosides, there was also a linear relationship between the percentage of hydrolysis and incubation time in digesta from the caecum (vicine r =0.97, P <0.0001; convicine r = 0.99, P < 0.0001 and the large intestine (vicine r = 0.79, P < 0.02; convicine r = 0.66, P < 0.05) (Fig. 2). The second experiment demonstrated that in vitro hydrolysis of vicine and convicine in digesta from neomycin-treated rats, compared with digesta from untreated rats, was inhibited to a considerable degree. The percentage of hydrolysis of vicine and convicine ioo [90~"
_
2
8ol-
70 ¢0 •~>, 60 -~
}
3 4
50 "°
5() ZO I0 I
o
io
zo 30 40 5o Incubation time (rnin)
60
70
Fig. 2. Hydrolysis of vicine (1) and convicine (2) in digesta from the caecum, and of vicine (3) and convicine (4) in
digesta from the large intestine over time. The regression equations are: (l) Y=4.825+I.172 X, ~2=0.951; (2) Y=5.518+ 1.293 X, ~2=0.973; (3) Y = 9.028 +0.934 X, ~2 =0.622; and (4) Y =21.359+0.641 X, ~2 =0.431.
838
M.S.S. ARBII)et aL Table I. Clearance of vicine and convicine in neomycin-treated and untreated rats determined in experiments 3 and 4 Urine (mg/24 hr) Treatment
Vicine
Convicine
Faeces (mg/24 hr) Vicine
Convicine
0.1 + 0 26.0+ 12
0.0 0.4+0.1
43.0 _+ 7.0 56.0_+21.0
22.0 _+ 11.0 82.0_+17.0
1.0 _+ 1.0 23.0 _+ 10.0
0.3 + 0.1 6.0 _+ 1.0
Experiment 3* Dietary vicine and convicine Control 1.0 ± 0.1 Treated 1.0+ 1.0
0.4 + 0.1 0.1 _+0.1 Experiment 4t Oral administration of vicine and convicine Control 3.0 _+ 2.0 1.0 _+ 1.0 Treated 12.0_+4.0 0.1 _ + 0 . 1 lp administration of vicine and eonvicine Control 5.0 _+ 2.0 5.0 _+ 1.0 Treated 42.0 _+ 16.0, 11.0 _+6.0
Values are means + SE (n = 5 rats per treatment). *Urine and faecal samples were collected for 48 hr after feeding of the diet containing 1% vicine and 1% convicine. AnalYSiS of variance showed that there was a significant (P < 0.05) three-way interaction between neomycin dose (treated or untreated), type of toxin (vicine or convicine) and route of clearance (urine or faeces). The total intake of vicine and convicine over 24hr was l l 0 m g for each of the two compounds, tUrine and faecal samples were collected 24 hr after oral or ip administration of the toxins. Analysis of variance showed that there were two significant (P < 0.05) three-way interactions: the first between neomycin treatment, type of toxin and route of toxin administration (oral or ip); and the second between type of toxin, route of administration and route of clearance. The total amount of vicine and convicine administered orally by intubation or ip was 750 and 600 mg/rat, respectively.
( + SE) in caecal digesta from control rats were 85 _ 2 and 100 + 1, respectively, while the values for caecal digesta from neomycin-treated rats were 29 + 1 and 81.1 ___1. For digesta from the large intestine, the percentage of hydrolysis of vicine and convicine was 88 + 8 and 78 + 1, respectively, in control digesta and 23 + 2 and 41 + 3, respectively, in digesta from neomycin-treated rats. Clearance of vicine and convicine in neomycin-treated and untreated rats The third experiment involved feeding of vicine and convicine to neomycin-treated and untreated rats. Assays of urine and faeces for the two glycosides showed that the amount of vicine in the faeces of neomycin-treated rats was much greater than the amount of vicine, or vicine plus convicine, in the urine and faeces of rats in all the other treatment groups (26mg v. < 2 m g , P < 0 . 0 5 ; Table 1). A maximum of 24% of the vicine and convicine that were consumed during the 24-hr period was excreted. Administration of vicine and convicine by intubation to neomycin-treated and untreated rats in experiment 4 also produced a glycoside-excretion pattern similar, but not identical, to that observed in experiment 3 (Table 1). In this experiment, as in the previous one, very little convicine was excreted in urine. The amount of vicine excreted in urine, however, was much higher in neomycin-treated rats than in untreated rats (12 v. 3 mg, P < 0.05). In faeces, the total excretion of convicine and that of vicine plus convicine were also much higher in neomycin-treated than in untreated rats (82 v. 22mg for convicine, P < 0 . 0 5 ; 138 v. 65mg for convicine plus vicine, P < 0.05).
The effect of neomycin on the excretion of vicine and convicine administered ip was also similar to that obtained when the glycosides were administered by intubation. The total amounts of vicine and convicine excreted in the urine of neomycin-treated and untreated rats were 53 and 10 mg (P < 0.05), respectively. The corresponding faecal values were 29 and 1.3 mg. Total excretion of convicine in urine, with and without neomycin treatment, was much greater in rats that received the glycosides ip compared with those given the toxins orally (16 v. 1.1 mg, P < 0.05). An opposite pattern was obtained in faeces (6.3 v. 104 mg, P < 0.05). Effect o f neomycin, vicine and convicine on blood glutathione concentration and mortality Analysis of variance carried out on the data from experiment 4 demonstrated that there was no effect related to the type of toxin (vicine or convicine, P < 0.05) on blood GSH concentration, but that there was an effect related to the route of administration of the toxins (P < 0.01). This interaction can be attributed to a much greater effect of the toxins on blood GSH concentration in rats that received the glycosides ip compared with those that received the toxins orally (12 v. 22 rag/100 ml). The effect of ip injection was particularly noticeable in rats that did not receive neomycin compared with those that received the antibiotic (12 v. 20 mg/100 ml). The corresponding GSH values for rats administered vicine and convicine orally were 22 and 23mg/100ml. Control rats administered physiological saline by either route had a GSH value of 28 _+ 1 rag/100 ml. In a similar manner, vicine and convicine were only toxic when administered ip (30% mortality), with the
Vicine and convicine hydrolysis in rats effect being much greater when neomycin was not provided in the diet (80% mortality). None of the control rats or those treated orally with the toxins died. DISCUSSION The /n vitro results demonstrated that vicine and convicine were hydrolysed more rapidly by digesta from the caecum and large intestine than by digesta from the small intestine or the stomach, and that the rate of hydrolysis was partly reduced when digesta was obtained from rats that had been fed diets containing neomycin. These observations are in agreement with those of Hegazy and Marquardt (1984). It would appear that anaerobic bacteria in the caecum and large intestine are primarily responsible for the hydrolysis of vicine and convicine. The present study also demonstrated that a considerable amount of vicine and convicine was hydrolysed in vivo and that neomycin increased greatly the amount of toxins excreted in urine and faeces, when the compounds were administered ip, and in faeces, when the compounds were given orally. The toxins administered ip were presumably secreted into the gastro-intestinal tract in a manner similar to that described by Arbid and Marquardt (1986). These researchers showed that vicine and convicine administered ip were rapidly transported into the large intestine where they were hydrolysed to their aglycones and then absorbed into the systemic system. Vicine and convicine administration decreased GSH concentration in blood, particularly in that of rats that were administered the toxins ip; this effect was associated with an increased mortality. Similar results were also obtained by Arbid and Marquardt (1988). These researchers concluded that, compared with the oral route, ip administration of the toxins was a better means of increasing the concentration of -i'hi~..glycosides and their subsequent hydrolytic products i~ ~the gastro-intestinal tract above the critical level required to produce a lethal effect. However, they did not orovide evidence that the lethal effect was attributable to the hydrolysis of the glycosides by the intestinal microflora. Further, the results of the present study demonstrated that neomycin pretreatment could reverse partially the decrease in blood GSH and reduce greatly the associated mortality in the group treated with the glycosides by the ip route. It may be concluded that vicine and convicine injected ip were transported rapidly into the large intestine and caecum and hydrolysed by digesta into their corresponding aglycones. The aglycones, when absorbed, led to a depletion of blood GSH and the death of the rats. Pretreatment with the poorly absorbed broad-spectrum antibiotic neomycin resulted in a reduced degree of hydrolysis of vicine and convicine and, consequently, to reduced effects on blood GSH and mortality. These observations support the hypothesis, and provide for the first time
839
direct evidence, that hydrolysis of the glycosides by micro-organisms in the gastro-intestinal tract must occur before these compounds become toxic. Acknowledgements--This research was supported by grants from the Natural Sciencesand Research Council of Canada, the Manitoba Department of Agriculture and the University of Manitoba. Excellenttechnical assistance was provided by Mrs Aniko Bematsky. REFERENCES
Albano E., Tomasi A., Mannuzzu L. and Arese P. (1984) Detection of a free radical intermediate from divicine of Vicia faba. Biochemical Pharmacology 33, 1701-1704. Arbid M. S. S. and Marquardt R. R. (1986) Effect of intraperitoneaUyinjected vicineand convicineon the rats: induction of favism-like signs. Journal of the Science of Food and Agriculture 37, 539-547. Arbid, M. S. S. and Marquardt, R. R. (1988) Favism-like effects of divicine and isouramil in the rat: acute and chronic effects on animal health, mortalities, blood parameters and ability to exchange respiratory gases. Journal of the Science of Food and Agriculture 43, 75-90. Belsey M. A. (1973) The epidemiologyof favism. Bulletin of the Worm Health Organization 48, 1-13. Beutler E., Duron O. and Kelly B. M. (1963) Improved method for the determination of blood glutathione. Journal of Laboratory Clinical Medicine 61, 882-888. Beutler E. (1978) Haemolytic Anaemia in Disorders of Red Cell Metabolism. pp. 96-101. Plenum, New York. Bottini E. (1973) Favism: current problems and investigations. Journal of Medical Genetics 10, 154-157. D'Aquino M., Gaetani S. and Spadoni M. A. (1983) Effects of factors of favism on the protein and lipid components of rat erythrocyte membrane. Biochimica et Biophysica Acta 731, 161-167. Donoso G., Hedayat H. and Khayatian H. (1969) Favism, with specialreference to Iran. Bulletin of the Worm Health Organization 40, 513-519. Frohlich A. A. and Marquardt R. R. (1983) Turnover and hydrolysis of vicine and convicine in avian tissues and digesta. Journal of the Science of Food and Agriculture 34, 153-163. Graf M., Frei B., Winterhalter K. H. and Richter C. (1985) Divicine induces calcium release from rat liver mitochondria. Biochemical and Biophysical Research Communications 129, 18-25. Hegazy M. I. and Marquardt R. R. (1984) Metabolism of vicine and convicine in rat tissue: absorption and excretion pattern and sites of hydrolysis. Journal of the Science of Food and Agriculture 35, 139-146. Halpcrin B. N., Beezi G., Mene G. and BenaceroffB. (1951) Etude quantitative de l'activit~ granulo-plexique du syst~me r~ticuloendoth~lial par l'injection intraveineux d'encre de Chine chez les diverses esp~ees animales. Annales de I'Institut Pasteur 80, 582-604. McBurney M. I. and Sauer W. C. (1993) Fiber and large bowel energy absorption: validation of the integrated ileostomy-fermentation model using pigs. Journal of Nutrition 123, 721-727. Mager J., Razin A. and Hershko A. (1969) Favism. In Toxic Constituents of Plant Foodstuffs. Edited by I. E. Liener. pp. 293-318. Academic Press, London. Marquardt R. R. (1989) Glycosides. In Toxicants of Plant Origin. Edited by P. R. Cheeke. Vol. 2. pp. 161-200. CRC Press, Boca Raton, FL. Marquardt R. R. and Frohlich A. A. (1981) Rapid-reverse phase high-performance liquid chromatographic method for the quantitation of vicine, convicine and related compounds. Journal of Chromatography 208, 373-379. Marquardt R. R., Muduuli D. S. and Frohlich A. A. (1983)
840
M.S.S. ARmD et al.
Purification and some properties of vicine and convicine isolated from faba bean (Vicia faba L.) protein concentrate. Journal of Agricultural and Food Chemistry 31, 839-844. Muduuli D. S., Marquardt R. R. and Guenter W. (1982) Effect of dietary vicine and vitamin E supplementation on productive performance of growing and laying chickens. British Journal of Nutrition 47~ 53-60. Rocic B., Rocic S., Ashcroft J. H., Harrison D. E. and Poje M. (1985) Diabetogenic action of alloxan-like compounds: the cytotoxic effect of dehydroisouramil hydro-
chloride on the rat pancreatic B-cells. Diabetologia Croatica 14, 143-149. Statistical Analysis System Institute Inc. (1986) SAS User's Guide: Statistics. 1986 Ed. pp. 584. SAS Institute Inc., Carey, NC. Yannai S. and Marquardt R. R. (1985) Induction of favismlike symptoms in the rat: effects of vicine and divicine in normal and buthionine sulphoximine-treated rats. Journal of the Science of Food and Agriculture 36, 1161-1168. Yoshida A. (1973) Haemolytic anaemia and G6PD deficiency. Science 179, 532-534.
0278-6915/93$6.00+ 0.00 Copyright © 1993PergamonPress Ltd
Printed in Great Britain.All rights reserved
EFFECT OF NEOMYCIN ON THE HYDROLYSIS AND TOXICITY OF VICINE AND CONVICINE IN RATS M. S. S. ARBID,M. S. MADHYASTHA,R. R. MARQUARDT*and A. A. FROHLICH Department of Animal Science, The University of Manitoba, Winnipeg, Canada R3T 2N2 (Accepted 11 June 1993)
Abstract--This study in the rat established the effects that a broad-spectrum and poorly absorbed antibiotic, neomycin sulfate, had on the in vitro and in vivo hydrolysis of vicine and convicine by the intestinal microflora, and on vicine- and convicine-induced depletion of blood glutathione and the resulting toxicity. The in vitro studies demonstrated that digesta from the caecum and large intestine were highly effective in hydrolysing vicine and convicine, whereas digesta from the same sections of the gastro-intestinal tract of neomycin-treated rats were much less effective (P < 0.0001). The in vivo studies showed that the total amount of vicine and convicine excreted in the urine and faeces was much greater in neomycin-treated rats compared with controls (P < 0.05), indicating the ability of neomycin to increase the amount of glycosides, particularly that of vicine, excreted in the faeces. The ability of glycosides to decrease the concentration of glutathione in blood (P < 0.05) and to increase rat mortality was greatly reduced in rats that were treated with neomycin, particularly in those treated ip with the toxin. Thus, the results demonstrated that neomycin reduced the rate at which vicine and convicine were hydrolysed in the lower section of the gastro-intestinal tract, and that neomycin treatment was associated with a reduced toxicity of the glycosides.
INTRODUCTION Vicine and convicine are glycosides that are found primarily in the faba bean (Vicia f a b a L.), which is one of the most important pulse crops in the world (Marquardt, 1989). Vicine and convicine are hydrolysed by the intestinal microflora (Frohlich and Marquardt, 1983; Hegazy and Marquardt, 1984) to the highly reactive free-radical-generating compounds (Albano et al., 1984) divicine and iso-uramil. Divicine and iso-uramil have been strongly suggested to be the causative agents in favism (Belsey, 1973; Bottini, 1973; Donoso et al., 1969), a haemolytic disease in humans, particularly in young males, that have a deficiency in erythrocytic glucose-6-phosphate dehydrogenase activity (Beutler, 1978; Yoshida, 1973). These free-radical generators may also cause other adverse effects including lipid peroxidation (D'Aquino et al., 1983), altered fat (Muduuli et al., 1982) mitochondrial metabolism (Graf et al., 1985), glutathione oxidation (Mager et al., 1969) and possibly diabetes (Rocic et al., 1985). Frohlich and Marquardt (1983) showed that a small percentage of dietary vicine, but not convicine was absorbed from the gastro-intestinal tract and excreted through the bile and kidney in chickens. Further, they demonstrated that the glycosides were not hydrolysed in vitro by avian tissue homogenates but were rapidly hydrolysed by digesta from the caecum of the birds. The total faecal excretion of
*To whom correspondence should be addressed. Abbreviation: GSH = glutathione.
vicine and convicine was increased when antibiotics were added to the diet. Similar results were obtained with the rat (Hegazy and Marquardt, 1984). In other studies, Arbid and M arquardt (1986) and Yannai and Marquardt (1985) demonstrated in rats that the route of administration of vicine and convicine greatly affected their toxicities. Oral administration of relatively large amounts of the glycosides to rats produced only small reductions in blood glutathione (GSH) concentrations and did not cause mortality. In contrast, ip administration of these compounds to rats resulted in a rapid decrease in the concentration of GSH in blood followed by death caused by anoxia. Rats that died from the toxic effects of these compounds had a large proportion of erythrocytes that were not able to exchange respiratory gases (Arbid and Marquardt, 1986 and 1988). These results cannot, however, be attributed to the direct hydrolysis of vicine and convicine by animal tissues (Frohlich and Marquardt, 1983; Hegazy and Marquardt, 1984) or to a direct effect of vicine and convicine on erythrocytes (Arbid and Marquardt, 1986). The results of the above studies indicated that, compared with the oral route, ip administration of vicine and convicine to rats resulted in a rapid uptake and hydrolysis of these compounds in the gastrointestinal tract, followed by the absorption of the aglycone into blood in sufficient quantities to overwhelm the protective effects of the GSH reductase system. Data from the literature therefore provide indirect (and not direct) evidence that vicine and convicine must first be hydrolysed to their aglycones to become toxic.
835
836
M.S.S. ARBIOet al.
The current study was based on the hypothesis that the toxicity of vicine and convicine administered ip to rats that are treated orally with neomycin will be decreased since the antibiotic will inhibit the growth of anaerobic micro-organisms in the gastro-intestinal tract, and consequently the decreased production of the aglycones divicine and iso-uramil will in turn result in a less marked effect on blood GSH concentration and lower mortality. The purpose of this study was to provide more definitive evidence to support this hypothesis. In order to achieve this goal, vicine and convicine were administered by both the ip and oral routes to control and neomycin-treated rats, and the mortality, blood GSH concentration and degree of hydrolysis of the glycosides by digesta from the gastro-intestinal tract were monitored. MATERIALS AND METHODS
Animals and treatment
Sprague-Dawley male albino rats (150_+5g) used in in vivo studies were obtained from the animal house of the National Research Centre (Dokki, Cairo, Egypt). Sprague-Dawley male albino rats (180-250 g) used in in vitro studies were obtained from the breeding farm of the University of Manitoba (Winnipeg, MB, Canada). The rats were housed individually in metabolic cages (Maryland Plastics, Inc., Federalsburg, MD, USA) and were maintained on a standard diet and given water ad lib. Chemicals
Vicine and convicine were prepared as described by Marquardt et al. (1983). Neomycin sulfate tablets (500 g each) were from Memphis Co. (Cairo, Egypt). Neomycin sulfate (powder) and fl-glucosidase were obtained from Sigma Co. (St Louis, MO, USA). Other chemicals were from Fisher Scientific Ltd (Winnipeg, MB, Canada). Estimation of vicine and convicine in urine, faeces and digesta
1 ml suspension of digesta or faeces in 0.145M saline (1 g/10 ml) was mixed with 5 ml perchloric acid (6%) and vortexed for 1 rain. Samples were passed through 0.5 ~m-filter for the clean-up. Vicine and convicine concentrations in these samples were measured by HPLC using a modification of the procedure described by Marquardt and Frohlich (1981). A Waters 600E Multisolvent Delivery HPLC System (Canada Waters Chromatography Division, ONT, Canada) with a 20-/~1 loop injector, a reversephase column (type Ultremex 5C18, 25 cm x 4.6 mm i.d., 5/~m; Phenomenex, Rancho Palos Verdes, CA, USA) and a variable-wavelength UV detector (Waters 484; Waters Associates Milford, MA, USA) set to 282nm and 0.01 units of sensitivity (absorbance units full scale or AUFS) was used. Glassdistilled water was used in the isocratic mobile phase in place of the eluting solution used by Marquardt
and Frohlich (1981). Divicine and iso-uramil were monitored in the absence of oxygen at 280 nm at a low pH (Marquardt and Frohlich, 1981). In vitro studies Hydrolysis of vicine and convicine in digesta. The first experiment used three rats weighing about 250 g. The animals were killed and digesta from the stomach, small intestine, caecum and large intestine were removed from each rat separately and weighed. Samples of digesta (! g each) from different parts of the gastro-intestinal tract were transferred separately and in duplicate from each of the three rats into capped test-tubes containing 10 ml 0.145 M degassed saline. The mixture was vortexed for i min, after which half of the volume in each tube was withdrawn and an equal volume of 0.145M degassed saline containing vicine (!.5 mg) and convicine (0.75 mg) was added. The mixture was incubated at 37°C for 10, 20, 40 and 60 min in a shaking water-bath. In the above studies, the saline solution was degassed by bubbling with oxygen-free nitrogen for 10 min. Incubation of digesta with vicine and convicine resulted in the maintenance of anaerobic conditions for at least 60min, and probably longer, as indicated by the persistence of the hydrolytic compounds divicine and iso-uramil, which have been shown to be readily inactivated by the presence of a trace amount of dissolved oxygen (Hegazy and Marquardt, 1984; Marquardt, 1989). Similar anaerobic in vitro fermentation procedures have been used to monitor the hydrolysis of dietary fibre by ileal digesta from pigs (McBurney and Sauer, 1993). In a second experiment, six rats weighing about 180-240 g were divided into two groups of three each and were housed individually in metabolic cages. The first group (control) was fed a normal stock diet, and the second group was given the diet containing 5% neomycin. The rats were killed after 3 days of feeding, and digesta from the caecum and large intestine of each rat (1 g each) were incubated with vicine and convicine for 30min as described above. In vivo studies Effect of neomycin on the hydrolysis of vicine and convicine. In the third experiment 10 male albino Sprague-Dawley rats weighing 150 + 5 g were divided into two groups of five each and were fed the standard diet containing 1% vicine and 1% convicine. The first group served as the control and the second group was treated by intubation with neomycin (250 mg/kg body weight, four times a day) for 24 hr prior to and during the collection period. Urine and faecal samples were collected over the following 48-hr period, and the pooled samples were assayed for vicine and convicine. Effect of neomycin on the toxicity of vicine and convicine administered orally or ip to rats. In the fourth experiment, 50 rats weighing 150 _+ 5 g were divided into 10 groups of five rats each. The 10
Vicine and convicine hydrolysis in rats B
A
RESULTS
C
HPLC assay for vicine and convicine
3
OJ
3
2
"6 O) ¢.) C
!
I
I
I
I
837
I
I
0 I 23456
!
I
!
I
0 12345
I
I
I
I
I
I
I
I
I
I
601254567
Retention time (min)
Fig. 1. Chromatograms of vicine and convicine incubation mixtures. The hydrolysis and elution patterns of vicine and convicinewere obtained when the two toxins were incubated with digesta from the caecum for different periods: 0.0 min (A), 30 rain (B) and 60 min (C). The different peaks correspond to the solvent (I), convicine (2) and vicine (3). experimental treatments were as follows: (1) control, physiological saline administered orally; (2) control, physiological saline administered ip; (3) control, vicine administered orally; (4) control, vicine administered ip; (5) control, convicine administered orally; (6) control, convicine administered ip; (7) neomycin administered orally followed by vicine given orally; (8) neomycin administered orally followed by vicine administered ip; (9) neomycin administered orally followed by convicine given orally; and (10) neomycin administered orally followed by convicine administered ip. Neomycin was administered to rats in groups 7 to 10 for 24 hr prior to administration of the glycosides and then throughout the 48-hr experimental period (250mg neomycin/kg body weight, four times a day). In all cases administration of the drugs was by intubation or ip injection. Vicine (5 g/kg body weight) and convicine (4g/kg body weight) were administered separately as indicated above. Rats that died in the groups treated ip with vicine and convicine (groups 4, 6, 8 and 10) were replaced with rats from the same stock colony. The same treatments were repeated with a sufficient number of rats so that the final total number of rats that survived for 48 hr was five per group. Blood samples from the orbital sinus of rats were collected in heparinized sterile tubes on ice, as described by Helperin et al. (1951), 2 hr after administration of vicine and convicine. Glutathione in blood was determined according to the procedure of Beutler (1963). Urine and faecal samples were collected for 24 hr after administration of the glycosides and were assayed for vicine and convicine as described above. Mortality represented the percentage of rats that died within 24 hr after vicine and convicine administration. Statistical analysis Data were subjected to analysis of variance (ANOVA), using the General Linear Model (GLM) procedure, and to regression analysis (Statistical Analysis System Institute Inc., 1986).
The assay procedure, which is a modification of that reported by Marquardt and Frohlich (1981), involved the elution of vicine and convicine in water instead of 0.05 M phosphoric acid-ammonia buffer (pH 2.0). The modified procedure improved the resolution of the two compounds, simplified the column clean-up and yielded a linear relationship between the peak areas of both vicine and convicine and the concentrations of the compounds used (i.e. up to 55/~g/ml for vicine and 25 #g/ml for convicine). A typical chromatogram for the hydrolysis of vicine and convicine is shown in Fig. 1.
In vitro hydrolysis of vicine and convicine The results of the first experiment demonstrated that the caecum and large intestine were the main sites for the hydrolysis of the two glycosides. The percentages of hydrolysis (_+ SE) of vicine and convicine, respectively, by digesta from different sections of the gastro-intestinal tract during a 60-min incubation period (at 37°C) were: 13.8 _+2.1 and 6.2 _+ 1.8 for the stomach, 27.8 _+ 1.2 and 45.4_+ 0.3 for the small intestine, 65.7 _+23.8 and 53.8 +_ 3.2 for the large intestine, and 77.7 + 0.1 and 85.25 _+0.25 for the caecum. For both glycosides, there was also a linear relationship between the percentage of hydrolysis and incubation time in digesta from the caecum (vicine r =0.97, P <0.0001; convicine r = 0.99, P < 0.0001 and the large intestine (vicine r = 0.79, P < 0.02; convicine r = 0.66, P < 0.05) (Fig. 2). The second experiment demonstrated that in vitro hydrolysis of vicine and convicine in digesta from neomycin-treated rats, compared with digesta from untreated rats, was inhibited to a considerable degree. The percentage of hydrolysis of vicine and convicine ioo [90~"
_
2
8ol-
70 ¢0 •~>, 60 -~
}
3 4
50 "°
5() ZO I0 I
o
io
zo 30 40 5o Incubation time (rnin)
60
70
Fig. 2. Hydrolysis of vicine (1) and convicine (2) in digesta from the caecum, and of vicine (3) and convicine (4) in
digesta from the large intestine over time. The regression equations are: (l) Y=4.825+I.172 X, ~2=0.951; (2) Y=5.518+ 1.293 X, ~2=0.973; (3) Y = 9.028 +0.934 X, ~2 =0.622; and (4) Y =21.359+0.641 X, ~2 =0.431.
838
M.S.S. ARBII)et aL Table I. Clearance of vicine and convicine in neomycin-treated and untreated rats determined in experiments 3 and 4 Urine (mg/24 hr) Treatment
Vicine
Convicine
Faeces (mg/24 hr) Vicine
Convicine
0.1 + 0 26.0+ 12
0.0 0.4+0.1
43.0 _+ 7.0 56.0_+21.0
22.0 _+ 11.0 82.0_+17.0
1.0 _+ 1.0 23.0 _+ 10.0
0.3 + 0.1 6.0 _+ 1.0
Experiment 3* Dietary vicine and convicine Control 1.0 ± 0.1 Treated 1.0+ 1.0
0.4 + 0.1 0.1 _+0.1 Experiment 4t Oral administration of vicine and convicine Control 3.0 _+ 2.0 1.0 _+ 1.0 Treated 12.0_+4.0 0.1 _ + 0 . 1 lp administration of vicine and eonvicine Control 5.0 _+ 2.0 5.0 _+ 1.0 Treated 42.0 _+ 16.0, 11.0 _+6.0
Values are means + SE (n = 5 rats per treatment). *Urine and faecal samples were collected for 48 hr after feeding of the diet containing 1% vicine and 1% convicine. AnalYSiS of variance showed that there was a significant (P < 0.05) three-way interaction between neomycin dose (treated or untreated), type of toxin (vicine or convicine) and route of clearance (urine or faeces). The total intake of vicine and convicine over 24hr was l l 0 m g for each of the two compounds, tUrine and faecal samples were collected 24 hr after oral or ip administration of the toxins. Analysis of variance showed that there were two significant (P < 0.05) three-way interactions: the first between neomycin treatment, type of toxin and route of toxin administration (oral or ip); and the second between type of toxin, route of administration and route of clearance. The total amount of vicine and convicine administered orally by intubation or ip was 750 and 600 mg/rat, respectively.
( + SE) in caecal digesta from control rats were 85 _ 2 and 100 + 1, respectively, while the values for caecal digesta from neomycin-treated rats were 29 + 1 and 81.1 ___1. For digesta from the large intestine, the percentage of hydrolysis of vicine and convicine was 88 + 8 and 78 + 1, respectively, in control digesta and 23 + 2 and 41 + 3, respectively, in digesta from neomycin-treated rats. Clearance of vicine and convicine in neomycin-treated and untreated rats The third experiment involved feeding of vicine and convicine to neomycin-treated and untreated rats. Assays of urine and faeces for the two glycosides showed that the amount of vicine in the faeces of neomycin-treated rats was much greater than the amount of vicine, or vicine plus convicine, in the urine and faeces of rats in all the other treatment groups (26mg v. < 2 m g , P < 0 . 0 5 ; Table 1). A maximum of 24% of the vicine and convicine that were consumed during the 24-hr period was excreted. Administration of vicine and convicine by intubation to neomycin-treated and untreated rats in experiment 4 also produced a glycoside-excretion pattern similar, but not identical, to that observed in experiment 3 (Table 1). In this experiment, as in the previous one, very little convicine was excreted in urine. The amount of vicine excreted in urine, however, was much higher in neomycin-treated rats than in untreated rats (12 v. 3 mg, P < 0.05). In faeces, the total excretion of convicine and that of vicine plus convicine were also much higher in neomycin-treated than in untreated rats (82 v. 22mg for convicine, P < 0 . 0 5 ; 138 v. 65mg for convicine plus vicine, P < 0.05).
The effect of neomycin on the excretion of vicine and convicine administered ip was also similar to that obtained when the glycosides were administered by intubation. The total amounts of vicine and convicine excreted in the urine of neomycin-treated and untreated rats were 53 and 10 mg (P < 0.05), respectively. The corresponding faecal values were 29 and 1.3 mg. Total excretion of convicine in urine, with and without neomycin treatment, was much greater in rats that received the glycosides ip compared with those given the toxins orally (16 v. 1.1 mg, P < 0.05). An opposite pattern was obtained in faeces (6.3 v. 104 mg, P < 0.05). Effect o f neomycin, vicine and convicine on blood glutathione concentration and mortality Analysis of variance carried out on the data from experiment 4 demonstrated that there was no effect related to the type of toxin (vicine or convicine, P < 0.05) on blood GSH concentration, but that there was an effect related to the route of administration of the toxins (P < 0.01). This interaction can be attributed to a much greater effect of the toxins on blood GSH concentration in rats that received the glycosides ip compared with those that received the toxins orally (12 v. 22 rag/100 ml). The effect of ip injection was particularly noticeable in rats that did not receive neomycin compared with those that received the antibiotic (12 v. 20 mg/100 ml). The corresponding GSH values for rats administered vicine and convicine orally were 22 and 23mg/100ml. Control rats administered physiological saline by either route had a GSH value of 28 _+ 1 rag/100 ml. In a similar manner, vicine and convicine were only toxic when administered ip (30% mortality), with the
Vicine and convicine hydrolysis in rats effect being much greater when neomycin was not provided in the diet (80% mortality). None of the control rats or those treated orally with the toxins died. DISCUSSION The /n vitro results demonstrated that vicine and convicine were hydrolysed more rapidly by digesta from the caecum and large intestine than by digesta from the small intestine or the stomach, and that the rate of hydrolysis was partly reduced when digesta was obtained from rats that had been fed diets containing neomycin. These observations are in agreement with those of Hegazy and Marquardt (1984). It would appear that anaerobic bacteria in the caecum and large intestine are primarily responsible for the hydrolysis of vicine and convicine. The present study also demonstrated that a considerable amount of vicine and convicine was hydrolysed in vivo and that neomycin increased greatly the amount of toxins excreted in urine and faeces, when the compounds were administered ip, and in faeces, when the compounds were given orally. The toxins administered ip were presumably secreted into the gastro-intestinal tract in a manner similar to that described by Arbid and Marquardt (1986). These researchers showed that vicine and convicine administered ip were rapidly transported into the large intestine where they were hydrolysed to their aglycones and then absorbed into the systemic system. Vicine and convicine administration decreased GSH concentration in blood, particularly in that of rats that were administered the toxins ip; this effect was associated with an increased mortality. Similar results were also obtained by Arbid and Marquardt (1988). These researchers concluded that, compared with the oral route, ip administration of the toxins was a better means of increasing the concentration of -i'hi~..glycosides and their subsequent hydrolytic products i~ ~the gastro-intestinal tract above the critical level required to produce a lethal effect. However, they did not orovide evidence that the lethal effect was attributable to the hydrolysis of the glycosides by the intestinal microflora. Further, the results of the present study demonstrated that neomycin pretreatment could reverse partially the decrease in blood GSH and reduce greatly the associated mortality in the group treated with the glycosides by the ip route. It may be concluded that vicine and convicine injected ip were transported rapidly into the large intestine and caecum and hydrolysed by digesta into their corresponding aglycones. The aglycones, when absorbed, led to a depletion of blood GSH and the death of the rats. Pretreatment with the poorly absorbed broad-spectrum antibiotic neomycin resulted in a reduced degree of hydrolysis of vicine and convicine and, consequently, to reduced effects on blood GSH and mortality. These observations support the hypothesis, and provide for the first time
839
direct evidence, that hydrolysis of the glycosides by micro-organisms in the gastro-intestinal tract must occur before these compounds become toxic. Acknowledgements--This research was supported by grants from the Natural Sciencesand Research Council of Canada, the Manitoba Department of Agriculture and the University of Manitoba. Excellenttechnical assistance was provided by Mrs Aniko Bematsky. REFERENCES
Albano E., Tomasi A., Mannuzzu L. and Arese P. (1984) Detection of a free radical intermediate from divicine of Vicia faba. Biochemical Pharmacology 33, 1701-1704. Arbid M. S. S. and Marquardt R. R. (1986) Effect of intraperitoneaUyinjected vicineand convicineon the rats: induction of favism-like signs. Journal of the Science of Food and Agriculture 37, 539-547. Arbid, M. S. S. and Marquardt, R. R. (1988) Favism-like effects of divicine and isouramil in the rat: acute and chronic effects on animal health, mortalities, blood parameters and ability to exchange respiratory gases. Journal of the Science of Food and Agriculture 43, 75-90. Belsey M. A. (1973) The epidemiologyof favism. Bulletin of the Worm Health Organization 48, 1-13. Beutler E., Duron O. and Kelly B. M. (1963) Improved method for the determination of blood glutathione. Journal of Laboratory Clinical Medicine 61, 882-888. Beutler E. (1978) Haemolytic Anaemia in Disorders of Red Cell Metabolism. pp. 96-101. Plenum, New York. Bottini E. (1973) Favism: current problems and investigations. Journal of Medical Genetics 10, 154-157. D'Aquino M., Gaetani S. and Spadoni M. A. (1983) Effects of factors of favism on the protein and lipid components of rat erythrocyte membrane. Biochimica et Biophysica Acta 731, 161-167. Donoso G., Hedayat H. and Khayatian H. (1969) Favism, with specialreference to Iran. Bulletin of the Worm Health Organization 40, 513-519. Frohlich A. A. and Marquardt R. R. (1983) Turnover and hydrolysis of vicine and convicine in avian tissues and digesta. Journal of the Science of Food and Agriculture 34, 153-163. Graf M., Frei B., Winterhalter K. H. and Richter C. (1985) Divicine induces calcium release from rat liver mitochondria. Biochemical and Biophysical Research Communications 129, 18-25. Hegazy M. I. and Marquardt R. R. (1984) Metabolism of vicine and convicine in rat tissue: absorption and excretion pattern and sites of hydrolysis. Journal of the Science of Food and Agriculture 35, 139-146. Halpcrin B. N., Beezi G., Mene G. and BenaceroffB. (1951) Etude quantitative de l'activit~ granulo-plexique du syst~me r~ticuloendoth~lial par l'injection intraveineux d'encre de Chine chez les diverses esp~ees animales. Annales de I'Institut Pasteur 80, 582-604. McBurney M. I. and Sauer W. C. (1993) Fiber and large bowel energy absorption: validation of the integrated ileostomy-fermentation model using pigs. Journal of Nutrition 123, 721-727. Mager J., Razin A. and Hershko A. (1969) Favism. In Toxic Constituents of Plant Foodstuffs. Edited by I. E. Liener. pp. 293-318. Academic Press, London. Marquardt R. R. (1989) Glycosides. In Toxicants of Plant Origin. Edited by P. R. Cheeke. Vol. 2. pp. 161-200. CRC Press, Boca Raton, FL. Marquardt R. R. and Frohlich A. A. (1981) Rapid-reverse phase high-performance liquid chromatographic method for the quantitation of vicine, convicine and related compounds. Journal of Chromatography 208, 373-379. Marquardt R. R., Muduuli D. S. and Frohlich A. A. (1983)
840
M.S.S. ARmD et al.
Purification and some properties of vicine and convicine isolated from faba bean (Vicia faba L.) protein concentrate. Journal of Agricultural and Food Chemistry 31, 839-844. Muduuli D. S., Marquardt R. R. and Guenter W. (1982) Effect of dietary vicine and vitamin E supplementation on productive performance of growing and laying chickens. British Journal of Nutrition 47~ 53-60. Rocic B., Rocic S., Ashcroft J. H., Harrison D. E. and Poje M. (1985) Diabetogenic action of alloxan-like compounds: the cytotoxic effect of dehydroisouramil hydro-
chloride on the rat pancreatic B-cells. Diabetologia Croatica 14, 143-149. Statistical Analysis System Institute Inc. (1986) SAS User's Guide: Statistics. 1986 Ed. pp. 584. SAS Institute Inc., Carey, NC. Yannai S. and Marquardt R. R. (1985) Induction of favismlike symptoms in the rat: effects of vicine and divicine in normal and buthionine sulphoximine-treated rats. Journal of the Science of Food and Agriculture 36, 1161-1168. Yoshida A. (1973) Haemolytic anaemia and G6PD deficiency. Science 179, 532-534.