Toxicity of Glucosinolates and Their Enzymatic Decomposition Products to Caenorhabditis Elegans

Home , Glucosinolate, Sinigrin

Journal of Nematology 27(3):258-262. 1995. © The Society of Nematologists 1995. Toxicity of Glucosinolates and Their Enzymatic Decomposition Products to Caenorhabditis elegans

STEVEN G. DONKIN, 1 MARK A. EITEMAN, 2 AND PHILLIP L. WILLIAMS 1'3

Abstract: An aquatic 24-hour lethality test using Caenorhabditis elegans was used to assess toxicity of glucosinolates and their enzymatic breakdown products. In the absence of the enzyme thioglucosidase (myrosinase), allyl glucosinolate (sinigrin) was found to be nontoxic at all concentrations tested, while a freeze-dried, dialyzed water extract of Crambe abyssinica containing 26% 2-hydroxyl 3-butenyl glucosinolate (epi-progoitrin) had a 50% lethal concentration (LC50) of 18.5 g/liter. Addition of the enzyme increased the toxicity (LCs0 value) of sinigrin to 0.5 g/liter, but the enzyme had no effect on the toxicity of the C. abyssinica extract. Allyl isothiocyanate and allyl cyanide, two possible breakdown products of sinigrin, had an LC50 value of 0.04 g/liter and approximately 3 g/liter, respectively. Liquid chromatographic studies showed that a portion of the sinigrin decomposed into allyl isothiocyanate. The resuhs indicated that allyl isothiocyanate is nearly three orders of magnitude more toxic to C. elegans than the corresponding glncosinolate, suggesting isothiocyanate formation would im- prove nematode control from application of glucosinolates. Key words: Caenorhabditis elegans, Crambe abyssinica, enzyme, epi-progoitrin, glucosinolate, myrosinase, physiology, sinigrin, thioglucosidase.

Glucosinolates are naturally occurring position products or between decomposi- compounds found primarily in plants of tion products. the family Cruciferae, where they are An objective of this work was to quantify thought to serve as repellents to potential the toxicity to the free-living nematode pests (5,10). Enzymatic decomposition of Caenorhabditis elegans of a commercially glucosinolates may occur rapidly within available glucosinolate, allyl glucosinolate the tissues of plants when the tissues are (sinigrin), and of a glucosinolate-con- damaged, and it may occur by soil micro- taining extract of Crambe abyssinica. A sec- organisms when plant material decom- ond objective was to quantify the toxicity poses. Recent work suggests that the prod- of the enzymatic decomposition products ucts of this complex degradation, such as of these glucosinolates in order to deter- organic cyanides (i.e., nitriles) or isothiocy- mine which, if any, of the decomposition anates, are the actual causes of toxicity (2). products elicited the greatest toxicity. Although several studies examining the Caenorhabditis eIegans is an excellent test nematicidal potential of glucosinolates on invertebrate for the rapid toxicological cruciferon plant extracts have been per- bioassessment of a variety of chemicals in formed (9,14-19,21-23), these studies aquatic medium (26,29), soils (6-8), and on have not distinguished potential glucosino- agar plates (20,26-30). Its suitability for late toxicity from the toxicity of the decom- such work is enhanced by its ease of culture and maintenance in the lab, as well as the enormous body of knowledge that ex- Received for publication 7 February 1995. i Environmental Health Science, College of Agricultural ists about its basic biology (31). and Environmental Sciences, University of Georgia, Athens, In this study, the C. elegans aquatic 24- GA 30602-7610. Current address for S. G. Donkin: Sciences hour lethality test developed by Williams International, Inc., King Street Station, 1800 Diagonal Road, Suite 500, Alexandria, VA 22314-2808. and Dusenbery (29) was used to assess allyl 2 Department of Biological and Agricultural Engineering, glucosinolate (sinigrin) and ex- College of Agricultural and Environmental Sciences, Univer- Crambe sity of Georgia, Athens, GA 30602-4435. tracts in both the presence and absence Corresponding author. of the glucosinolate-hydrolyzing enzyme The authors express thanks to the North Dakota Experi- ment Station for a sample of Cramba abyssinica. The authors thioglucosidase (myrosinase). In addition, are grateful to the U.S. Department of Agriculture (USDA) this lethality test was used to assess the tox- High Erucic Acid Development Effort, the USDA Advanced Materials From Renewable Resources Program, the Univer- icity of pure samples of allyl isothiocyanate sity of Georgia's Agricultural Experiment Station and the and allyl cyanide (a nitrile). Georgia Power Company for financial support of this re- search. Studies such as this may lead to greater 258 Toxicity of Glucosinolates Using C. elegans: Donhin et al. 259 use of standardized C. elegans toxicity test tional tests were performed with sinigrin protocols as initial screens for assessing the (5, 10, and 20 g/liter) and enzyme (I Ixg/g potential of various compounds for pest of sinigrin) at pH conditions and a ferrous control. Although C. elegans is a free-living ion concentration adjusted to promote ei- species, it is known to be a potential pest to ther total allyl cyanide formation (pH 4.0, mushroom crops (13), and the evolution of 10 mM ferrous ion) or allyl isothiocyanate rhabditid nematodes may have included formation (pH 10.0, no ferrous ion) (2,5, other parasitic forms (24). Caenorhabditis el- 25), followed by the nematode test. Four egans also has the advantage of being much adult C. elegans were added to each well, easier to culture in the lab than most par- and mortality was assessed after 24 hours asitic nematodes. (28,29). Each test was replicated at least 10 times, and LC50 s (concentration resulting MATERIALS AND METHODS in 50% mortality) were determined for each compound by probit analysis (11). Cultures of wild type strain N2 Cae- In order to determine the actual break- norhabditis elegans were maintained in the down products of sinigrin with thioglucosi- dauerlarva state at 20 C in M9 buffer (3). dase obtained under the test conditions, an Two days before beginning a test, several analysis was conducted (without nema- hundred dauerlarvae were transferred to todes) to determine the composition of the a petri plate containing K-agar (28) with solution over the test period. The concen- a mature lawn of Escherichia coli strain trations of sinigrin, epi-progoitrin (a glu- OP50 (3) as a food source. After two days cosinolate) and allyl isothiocyanate in solu- at 20 C, the nematodes had developed into tion were quantified by liquid chromatog- synchronous adults that were then trans- raphy (1). ferred with a sterile platinum wire to aquatic test samples (29). RESULTS Test solutions consisted of the following In the absence of thioglucosidase, sini- commercially available compounds dis- grin was not toxic to C. elegans up to the solved in deionized water: sinigrin mono- greatest concentration tested (80 g/liter = hydrate (allyl glucosinolate; Sigma, St. 193 raM) (Fig. 1). In contrast, the addition Louis, MO), allyl cyanide (98% pure; Ald- of thioglucosidase resulted in an increased rich, Milwaukee, WI), and allyl isothiocy- toxicity of at least two orders of magnitude anate (94% pure; Janssen Chimica, Beerse, Belgium). Crambe abyssinica seed extract was prepared by blending seeds in a 5:1 ratio (100 C water:seed) (v/v) for 8 minutes ,0080...... "3 ...... followed by centrifugation at 500x g for I0 minutes. The supernatant was dialyzed o~ (12,000 molecular weight cutoff) for 24 v 60 hours against four volumes of water at 4 C .>__ and then freeze-dried to produce a solid 40 that was approximately 26.5% 2-hydroxyl co 3-butenyl glucosinolate. 20 Test solutions containing thioglucosidase (Sigma, enzyme activity of 200 units/ 0 i ...... ~ ......

g) were prepared as above, with the en- J,I , , , ,,,,,I , , ,,,,,,I , , ,,,,itl i i illJl,I zyme added to achieve a concentration of 1 0.01 0.10 1.00 10.00 100.00 ~xg/g of either sinigrin or Crambe seed ex- Concentration (g/L) tract. Tests were performed using freshly FIG. 1. Percentage of C. elegans survival after 24 made solutions in microtiter wells (Falcon hours of exposure to concentrations of sinigrin (O) 3047) with 0.5 ml solution in each. Addi- and sinigrin with thioglucosidase (0). 260 Journal of Nematology, Volume 27, No. 3, September 1995

(LCs0 = 0.5 g/liter = 1.2 raM). A control with enzyme alone was found to be non- 100 toxic. Exposure to Crambe seed extract re- 8O sulted in a mortality range between 5 and o~ 50 g/liter (LC50 = 18.5 g/liter) (Fig. 2). In _~" 60 contrast to the results using pure sinigrin, .>_ the toxicity was not significantly changed ~ 40 by addition of thioglucosidase. cO Under certain chemical conditions, thio- 20 glucosidase will convert glucosinolates into organic cyanides (i.e., nitriles) or isothiocy- 0 .~ ...... anates. Two possible products from the decomposition of sinigrin are allyl cyanide 0.01 0.10 1.00 10.00 100.00 and allyl isothiocyanate. Figure 3 shows Concentration (g/L) the concentration-response curves for these FIG. 3. Percentage of C. elegans survival after 24 two compounds after 24 hours using the C. hours of exposure to concentrations of allyl isothio- elegans toxicity test. Exposure to allyl cyanate (4)) and allyl cyanide (O). isothiocyanate at concentrations between 0.01 and 0.1 g/liter resulted in nematode cyanide was expected to predominate (low mortality (LC50 = 0.04 g/liter = 0.5 mM), pH, 10 mM ferrous ion) (data not shown). while allyl cyanide was less toxic, having an The concentration of isothiocyanate in LC50 of approximately 3 g/liter (45 raM). the sinigrin with thioglucosidase solution The sinigrin and thioglucosidase tests (at was found to be 12 to 31% of the gluco- concentrations of 5, 10, and 20 g/liter of sinolate concentration after 24 hours. sinigrin), which were incubated under However, allyl isothiocyanate (both pure known pH conditions and controlled fer- and produced by enzymatic decomposi- rous ion concentrations, were toxic where tion) was found to be chemically unstable allyl isothiocyanate was expected to pre- in the unbuffered aqueous solution, and dominate (high pH, no ferrous ion). No disappeared from solution within 1 week mortality occurred in solutions where allyl at4 C. The pH of the solutions was typically in the range of 5.5 to 6.0. An exception was 100 the solution with sinigrin and thioglucosidase, which after 24 hours had a pH of 4.0, 8O probably due to sulfate production by the reaction mixture (12). Fluctuations within "~ 60 this range of pH have been found to have ._> no effect on mortality in the C. elegans toxicity test (Donkin and Williams, unpubl. 40 Or) obs.).

20 DISCUSSION

0 ...... Several authors have speculated that I i ~ i i lille i i t , , ,,,I pesticidal effects from glucosinolates are 10 100 largely due to their enzymatic decomposi- Concentration (g/L) tion products rather than the glucosino- lares themselves (2,4,10,17,19). These de- FIG. 2. Percentage of C. elegans survival after 24 hours of exposure to concentrations of Crambe extract composition products consist mostly of al- (©) and Crambe extract with thioglucosidase (O). lyl isothiocyanate and allyl cyanide (i.e., Toxicity of Glucosinolates Using C. elegans: Donhin et al. 261 nitrile), with their relative production de- this result would be expected if, as the pending on pH, ferrous ion, and gluco- chromatographic evidence shows, sinigrin sinolate side-chain structure (2,25). Borek did not completely decompose to form this et al. (2) found that acidic pH and 10 mM product. Since the enzymatic conversion ferrous ion favor allyl cyanide production was found to be 12-31% isothiocyanate, from enzymatic hydrolysis of sinigrin, the toxic agent in the sinigrin-enzyme so- while allyl isothiocyanate is the principal lutions appears to be aUyl isothiocyanate. product at higher pH. Ferrous ion was This finding is further supported by the found to inhibit the formation of both testing of sinigrin with enzyme at pH con- products at pH 4 and 6 (2), although Chew ditions and a controlled ferrous ion con- (5) reports that ferrous ion may preferen- centration that found C. elegans mortality tially promote nitrile formation. occurred only in those solutions where al- Our results confirm that the enzymatic lyl isothiocyanate would be expected to decomposition products of sinigrin rather predominate. than sinigrin itself, are toxic. The low- Caenorhabditis elegans aquatic toxicity molecular-weight extract from Crambe ab- tests support results of other researchers yssinica seeds, which contains many com- who found glucosinolates to be relatively pounds other than glucosinolate, was toxic non-toxic to nematodes except when re- only at fairly high concentrations, and its acted with thioglucosidase (14-17,19). toxicity was unaffected by incubation with However, a breakdown product of some thioglucosidase. It is possible that the prin- glucosinolates--isothiocyanate--is rela- cipal glucosinolate in the Crambe extract, tively toxic to nematodes and might be use- epi-progoitrin, was hydrolyzed into ox- ful as an effective nematicide. azdidinethione rather than isothiocyanate (5). Therefore, some unknown chemical in LITERATURE CITED the extract not affected by the addition of 1. Bj6rkquist, B., and A. Hase. 1988. Separation thioglucosidase would appear to be the and determination of intact glucosinolates in rape- toxic agent in the Cram,be extract. seed by high-performance liquid chromatography. Based on the results of other studies (2, Journal of Chromatography 435:501-507. 2. Borek, V., M.J. Morra, P. D. Brown, and J. P. 5,25), the conditions for enzymatic reac- McCaffrey. 1994. Allelochemicals produced during tion of sinigrin in this study (pH >3.5, no sinigrin decomposition in soil. Journal of Agriculture ferrous ion) suggest that allyl isothiocy- and Food Chemistry 42:1030-1034. anate, rather than allyl cyanide, should be 3. Brenner, S. 1974. The genetics of Caenorhabditi~ elegans. Genetics 77:71-94. the primary decomposition product. Tests 4. Brown, P.D., M.J. Morra, J. P. McCaffrey, with pure allyl cyanide and allyl isothiocy- D.L. Auld, and L. Williams. 1991. Allelochemicals anate support this prediction. If sinigrin produced during glucosinolate degradation in soil. were to have reacted completely in the Journal of Chemical Ecology 17:2021-2034. 5. Chew, F. S. 1988. Biological effects of glucosino- presence of enzyme to form a stoichiomet- lates. Pp. 155-181 in H. G. Cutler, ed. Biologically ric amount of a product, then the concen- active natural products: Potential use in agriculture. tration-response curve for the sinigrin and Washington, D.C.: American Chemical Society. enzyme solution would resemble that of ei- 6. Donkin, S. G. 1993. A soil toxicity test using the nematode Caenorhabditis elegans and some applica- ther pure allyl cyanide or allyl isothiocy- tions to studying metal ion sorption processes in soils. anate, whichever was the primary product. Ph.D. thesis. Georgia Institute of Technology, At- Since pure allyl cyanide is less toxic than lanta. sinigrin with enzyme by 1.5 orders of mag- 7. Donkin, S. G., and D. B. Dusenbery. 1993. A soil toxicity test using the nematode Caenorhabditis elegans nitude, the primary reaction product can- and an effective method of recovery. Archives of En- not be allyl cyanide. Any deviation due to vironmental Contamination and Toxicology 25:145- incomplete sinigrin reaction would be less 151. toxic than the pure product. Pure allyl 8. Donkin, S. G., and D. B. Dusenbery. 1994. Us- ing the Caenorhabditis elegans soil toxicity test to iden- isothiocyanate is an order of magnitude tify factors affecting toxicity of four metal ions in in- more toxic than sinigrin with enzyme, and tact soil. Water, Air and Soil Pollution 78:359-373. 262 Journal of Nematolog),, Volume 27, No. 3, September 1995

9: Ellenby, C. 1945. Control of the potato-root endpoints. Ph.D. thesis. Georgia Institute of Technol- eelworm, Heterodera rostochiensis Wollenweber, by al- ogy, Atlanta. lylisothiocyanate, the mustard oil of B~assica nigra L. 21.Mojtahedi, H., G.S. Santo, A.N. Hang, and Annals of Applied Biology 32:237-239. J. H. Wilson. 1991. Suppression of root-knot nema- 10. Fenwick, G. R., R. K. Heaney, and W.J. Mul- tode populations with selected rapeseed cultivars as lin. 1983. Glucosinolates and their breakdown prod- green manure. Journal of Nematology 23:170-174. ucts in food and plants. CRC Critical Reviews in Food 22. Mojtahedi, H., G. S. Santo, J. H. Wilson, and Science and Nutrition 18: 123-201. A. N. Hang. 1993. Managing Meloidogyne chitwoodi on 11. Finney, D.J. 1971. Probit Analysis. New York, potato with rapeseed as green manure. Plant Disease NY: Cambridge University Press. 77:42-46. 12. Gil, V., and A. McLeod. 1980. The effects 23. Pinkerton, J. N., H. Mojtahedi, and G. Santo. of pH on glucosinolate degradation by a thiogluco- 1987. Reproductive efficiency of Pacific Northwest side glucohydrolase preparation. Phytochemistry 21: populations of Meloidogyne chitwoodi on alfalfa. Plant 2547-2551. Disease 71:345-348. 13. Grewal, P. S., and P. N. Richardson. 1991. Ef- 24. Poinar, G. O., Jr. 1983. The Natural History of fects of Caenorhabditis elegans (Nematoda: Rhabdidae) Nematodes. Englewood Cliffs, NJ: Prentice-Hall. on yield and quality of the cultivated mushroom Agar- 25. Uda, Y., T. Kurata, and N. Arakawa. 1986. Ef- icus bisporus. Annals of Applied Biology 118:181-194. fects of pH and ferrous ion on the degradation of 14. Jing, G. N., and J. M. Halbrendt. 1992. Evalu- glucosinolates by myrosinase. Agricultural and Bio- ation of rapeseed extract toxicity using a Caenorhab- logical Chemistry 50:2735-2740. ditis elegans bioassay. Journal of Nematology 24:600- 26. Williams, P. L. 1988. Evaluation of Caenorhab- 601 (Abstr.). ditis elegans as an acute lethality and a neurotoxicity 15. Jing, G. N., andJ. M. Halbrendt. 1993. Nema- screening model. Ph.D. thesis. Georgia Institute of totoxins in rapeseed. Journal of the Pennsylvania Technology, Atlanta. Academy of Science 66:168 (Abstr.). 27. Williarns, P.L., and D.B. Dusenbery. 1987. 16. Jing, G. N., and J. M. Halbrendt. 1994. Rela- Screening test for neurotoxins using Caenorhabditis el- tionship of glucosinolate in rapeseed extracts to tox- egans. Pp. 163-170 in A. Shahar and A. Goldberg, eds. icity against Caenorhabditis elegans. Journal of Nema- Model Systems in neurotoxicology: Alternative ap- tology 26:104 (Abstr.). proaches to animal testing. New York, NY: Alan R. 17. Jing, G. N., andJ. M. Halbrendt. 1994. Nema- Liss. ticidal compounds from rapeseed (Brassica napes and 28. Williams, P. L., and D.B. Dusenbery. 1988. B. campestris). Journal of the Pennsylu'ania Academy Using Caenorhabditis elegans to predict mammalian of Science 68:29-33. acute lethality to metallic salts. Toxicology and Indus- 18. Johnson, A. W., A. M. Golden, D. L. Auld, and trial Health 4:469-478. D. R. Summer. 1992. Effects of rapeseed and vetch as 29. Williams, P. L., and D. B. Dusenbery. 1990. green manure crops and fallow on nematodes and Aquatic toxicity testing using the nematode Cae- soilborne pathogens. Journal of Nematology 24:117- norhabditis elegans. Environmental Toxicology and 126. Chemistry 9:1285-1290. 19. Lazzeri, L., R. Tacconi, and S. Palmieri. 1993. 30. Williams, P. L., and D.B. Dusenbery. 1990. In vitro activity of some glucosinolates and their reac- Promising indicator of neurobehavioral toxicity using tion products toward a population of the nematode the nematode Caenorhabditis elegans and computer Heterodera schachtii. Journal of Agriculture and Food tracking. Toxicology and Industrial Health 6:425- Chemistry 41:825-829. 440. 20. Middendorf, P. 1994. Development and evalu- 31. Wood, W. B., ed. 1988. The nematode Cae- ation of toxicity tests using Caenorhabditis elegans with norhabditis elegans. Cold Spring Harbor, NY: Cold reproduction, mutation, lethality, and behavior as Spring Harbor Laboratory Press.