PHYTOCHEMICAL ANALYSIS, VOL. 3, 91-94 (1992)

ACCELERA TED COMMUNICA TlONRapid Screening Assay of Cyanide Content ofCassava

----------------1 'º~, I

f.f ,~\"'.. ~. .,~ [ '. , '"' - r-, '"~ ~ , .•..J;...•J J.; ':'I'/r.." 'I\:o' ." '.......1 : \Ut,l~..-:tf

J. Howard Bradbury* and Sylvia V. EganDivision of Botany and Zoology, Australian National University, Canberra, ACT, 2601, Australia

A rapid, inexpensive screening assay was developed in order to measure the cyanide content of cassava (Manihotesculenta Crantz) tubers. A small disc of parenchyma tissue cut with a cork borer or aIternatively grated tissuewas placed in a stoppered glass tube and the hydrogen cyanide Iiberated produced a blue colour on a filter paperpreviously spotted with a drop of tetra-base [4,4' ·methylenebis-(N,N-dimethylaniline)] and cupric acetate. Theintensity of the blue colou r which developed within one hour was rated visually on a graded scale from Oto S. Thecorrelation coefficient between samples accurately analysed for total cyanide and also tested using the rapid assaywas 0.77.

Keywords: Cyanide content; cassava; Manihot esculenta; linamarin; cyanogenic glycosides.

INTRODUCTION

Cassava (Manihot esculenta Crantz) is the third mostimportant food source in the tropical world after riceand maize and is the most important food in the humidand semi-humid tropics of Africa. Cassava tubers andleaves contain cyanogenic glucosides, mainly linamarinand a small amount of lotaustralin (Bradbury andHolloway, 1988). The cyanogenic glucosides and lina-marase are physically separated from one another inthe plant and disruption of the tissue allows the enzymelinamarase to come into contact with the former andcatalyse the hydrolysis. Thus linarnarin is hydrolysed toacetone cyanohydrin which further breaks down toacetone and hydrogen cyanide.

Hydrogen cyanide is a lethal poison at a levei of 0.5-3.5 mg HCN/kg body weight (Montgomery, 1980;Solomonson, 1981). The parenchyma of cassava tuberscontains cyanogenic glucosides equivalent to 0.2-100 mg HCNIlOO g fresh weight (Bradbury andHolloway, 1988; Bradbury et al., 1991), hence high-cyanide varieties of cassava present a considerablehazard to human health, even though post-harvesttreatrnents including cooking reduce the cyanide con-tent. As well as the possibility of acute intoxication anddeath, regular exposure to sublethal amounts of cya-nide may exacerbare goitre and cretinism (Delange etal., 1983) and cause tropical neuropathy (Osuntokun,1981) and epidemic spastic paraparesis (Rosling, 1987),which is also called konzo (Howlett et al., 1990).

A simple rapid screening method for cyanide incassava is important for cassava breeding programmes.One method used at the International Institute forTropical Agriculture (IIT A), utilizes sodium picratepaper that is suspended above a leaf disc in a sealedglass tube (Guignard, 1906; Sadik et al., 1974; Cooke etal., 1978). The HCN liberated causes a colour change.frorn yellow to red-brown.

* Author 10 whorn correspondence should be addressed.

0958-0344/92/020091-04 $05.00© 1992 by John Wiley & Sons. Lld.

One problem with this method is the poor correlationbetween the cyanide content of leaf and tuber. Cookeet ai. (1978) found a significant correlation (P<0.05)between picrate leaf score and root cyanide content(based on fresh weight of peeled root) for 108 clones(correlation coefficient r = 0.22). The correlation wasimproved when the root cyanide content was caJculatedon a dry weight basis (P about 0.01, r = 0.36). A similarcorrelation coefficient was obtained by Dr M.Bokanga by calculation of the data of de Bruijn (1971).On the other hand, Mkpong et al . (1990) found that theaverage linamarin content of leaves from low and highcyanide varieties of cassava was essentially the same. Asecond problem is that the picrate method is unreliableas a quantitative method for cyanide (Williams andEdwards, 1980; Fukuba and Mendoza, 1981;lzomkun-Etiobhio and Ugochukwu, 1984).

Another method (Feigl, 1954; Esquivil andMaravalhas, 1973) uses benzidine and copper acetate,but this was abandoned because of the carcinogeniccharacter of benzidine (Feigl and Anger, 1966). Thelatter authors replaced the benzidine spot test by tetra-base [4,4' -methylenebis-(N,N-dimethylaniline) 1 andcopper ethylacetoacetate in chloroform and thismethod was used by Tantisewe et aI. (1969),on plantmaterial. We have further developed this method usingtetra-base in acetone and cu ric acetate in acetic acid :Water as a rapid screening assay for cyanide in tubersand leaves.

EXPERIMENTAL

Materiais. Terra-base was obtained from Kodak (EastmanKodak , Rochester, NY) and recrystallized from ethanol. __Cupric acetate (laboratory grade reagent) was obtained fromMay and Baker , Dagenham, UK. Cupric ethylacetoacetate

Received 17 October 1991Accepted (revised) 12 December 1991

""" ..,. .•..,~.-Iil.1q

-

92 J. H. BRADBURY AND S. V. EGAN

was preparcd from acetoacetic acid ethyl ester (SigmaChemical Co.. St. Louis, USA) and cupric acetate (Vogel,1989).

Methods. Fresh cassava tubers and leaves were obtained fromplants grown in a glasshouse in Canberra. Tubers werewashed and air-dried. Because of the relatively smalllongitu-dinal gradient of linamarin concentration from the proximalto the distal end of the tuber, one circular section of thickness2-5 mm cut across the peeled midsection of the tuber wasused (Cooke, 1978; Bradbury et ai., 1991). Care was taken toremove the cortex layer. Since there is a much steepergradient of concentration of linamarin from the periphery(high) to the centre (Iow) of the tuber (Cooke, 1978; Kojimaet aI., 1983) two different methods of sampling the circularsections were used as follows:

(1) For small tubers of diameter up to about 40 mm areasonable sample was obtained by cutting a circular disc oftissue out of the section using a cork borer of diameter 21 mmplaced at the edge of the section.

(2) To obtain accurate sampling of the section cut from thetuber, either the whole section, or a suitably sized segmentcut from it, was grated and the grated material thoroughlymixed.

Either the 21 mm diameter disc (Method 1) or a spoonful ofthe grated material (Method 2) was placed in a glass tube ofinternal diameter about 23 mm and length 50 mm sealed witha plastic capo A rectangular strip of filter paper (about10 mm x 45 mm) was held between the cap and the glass tubewith about 35 mm hanging inside the tube. The filter paperhad previously been spotted near the end with one drop of theindicator solution (see below) and the solvent allowed toevaporate completely before insertion into the tube. Thisrequired not less than 30 min at 20°C; use of damp strips ofpaper gave slower colour developrnent. The indicator solu-tion (prepared freshly each day) was obtained by mixingequal volumes of 0.3% (w/v) tetra-base in acetone and 0.3%(w/v) cupric acetate in 15% acetic acid in water. Each ofthese 0.3% sol~tions were stable for months when keptseparately. Use of more dilute solutions (0.15%) causedslower colour development.

A blue colour developed and intensified over a period of 5-60 min at roorn temperature, depending on the amount ofHCN liberated from the cassava disco Six levels were dis-tinguished as follows-O: no colour (obtained with a blanktest containing no cassava disc), 1: very low intensity (verypale blue), 2: low (pale blue), 3: rnediurn blue, 4: moderatelyhigh, 5: high (blue-purple). Using sampling Method 1, inorder to distinguish between sarnples in the high range (10-100 mg HCNIlOOg fresh weight), the colour intensities werecompared after about 15 min at 20 °C, whilst for the low range«10 rng HCNIlOOg fresh weight), a time of 60 rnin at about20°C, or 15-30 min at 27°C, was required. In ali cases thecolour intensified with time and after 16 h differencesbetween different samples were largely eliminated. Earlytests were made with cupric ethylacetoacetate (Feigl andAnger , 1966), but cupric acetate was found to be equallyeffective and was used in ali subsequent work, because of itscheapness and ready availability. However, with cupric ace-tate, but not with cupric ethylacetoacetate, there is slowformation of a blue colour over about 16 h at 20°C, even inthe absence of HCN. When grated samples (Method 2) werecompared with discs (Method 1) using the same samp!e of

- cassava, it was found that the latter were norrnally 1-:-2levelslower in ,colour than the former. Thus it was important not tomix up the two methcds in the one series of experiments.

The total cyanide content of cassava .samples was alsoobtained using the recently developed acid hydrolysis methodof Bradbury et ai. (1991).

RESULTS AND DISCUSSION

According to Feigl and Anger (1966) the tetra-base isoxidized to a blue conjugated product by the copper(n), which is reduced to copper (I), in a reaction thatinvolves cyanide. The original technique used cuQIiç.eth lacetoacetate, which has been replaced in the pres-êrrt method by the more readily available and cheapercu ric acetate. Also chloroform has been replaced as asolvent by less toxic and cheaper acetone: acetic acid:water.

The blue colour intensified with time, due to thecontinuing liberation of HCN by the hydrolysis oflinamarin (catalysed by linamarase) and the breakdownof unstable acetone cyanohydrin to HCN and acetone.Thus, at any given time after starting the experiment,the intensity of colour depended on the rate of liber-ation of HCN, which in turn depended on (1) the areaof cut surface, (2) the temperature, (3) the amount oflinamarase and possibly hydroxynitrile lyase presentand (4) the amount of linamarin in the sarnple. Asalready mentioned, increasing the area of the cut sur-face by grating (Method 2) increased the colour inten-sity. A series of parenchyma samples made with discthicknesses of 1-8 mm showed that there was noappreciable difference in colour intensity for thick-nesses between 2 and 5 mm. Thus, parenchyrna discs ofthickness abouO mmwere used.

We have recently shown that the rate of autolysis oflinamarin in cassava parenchyma tis sue increased withincrease in linamarase content (Bradbury and Egan,1991), hence variation in the linamarase content (andpossibly also the amount of hydroxynitrile lyase) fromsample to sample must affect adversely the accuracy ofthe test method. The extent of this effect could perhapsbe ascertained by accurate measurement of the lina-marin and linamarase contents of tissue which was alsosubjected to the rapid assay. As shown in Table 1 wehave compared the linamarin content by the acidhydrolysis and rapid assay methods, but have not mea-sured the linamarase content. There is a reasonablygood correlation between the results of the rapid assayand the total cyanide content determined by acidhydrolysis (r = 0,77 for 16 samples containing from 0.8to 11 mg HCN/lOOg fresh weight).

Sampling of cassava plants for cyanide

The presently used picrate leaf test for cyanide (Sadiket ai., 1974; Cooke et al., 1978) is based on a correlationbetween leaf cyanide and tuber cyanide. Although asignificam degree of correláiion was claimed by Cookeet alo (1978) there is increasing evidence that the corre-lation is not good enough for the leaf method to give areliable indication of tuber cyanide content. Sinceplants cannot be sarnpled until they are mature enoughto produce., tubers, it is necessary to sample tuberparenchyma directly. The two methods of sampling thetubers proposed here should not be used indiscrimina-

-

RAPID ASSA Y OF CYANIDE CONTENT OF CASSA VA 93

Table 1. Comparison of total cyanide content of parenchyma and cortex tissueof cassava determined by acid hydrolysis and rapid screening assay

Cultivar"

SM1-150SM1-150SM1-150SM1-150SM1-150SM1-150SM 1-150 cortexSM 1-150 cortexM Col 1468M Càl 1468M Col 1468M Col 1468CeibaCeibaCeibaCeiba cortexMAus 7MAus 7

Acid hvdrolvsis"(mg HCN/l00g fresh wt.)

1.80.801.71.21.61.6

11233.85.05.58.62.64.05.2

502.83.5

Colour score by rapid assav''following an incubation time of

30 min 60 min

3-4

1333

1-23553

2-3352

3-4352

2-3

24

5

• Ali samples were of parenchyma tissue unless stated otherwise.b Quantitative method of Bradbury et aI. (1991).c Sampling was by Method 1.

tely in the one experirnent, because the rate of produc-tion of HCN frorn a grated sample (Method 2) is muchgreater than frorn a disc (Method 1) of the samesample. Method 2 gives a more representative samplethan Method 1, but the latter is faster and probablyadequate for use except with large tubers of diametergreater than about 50 mrn.

Comparison of the tetra-base method with the picratemethod

The picrate method is less sensitive than the pyridine-pyrazolone (Ikediobi et al., 1980) or the pyridine-barbituric acid methods. We have tested both methodson leaf discs (moistened with a drop of toluene; Sadik etal., 1974), cortex and parenchyma tissue discs. Withlow cyanide parenchyma discs the picrate method gaveonly a very slight colour change, whereas tetra-basegave good results, hence the latter is more sensitive, inagreement with Brinker and Seigler (1989). The tetra-base method gives more rapid colour development

(1 h) than the picrate method (4 h) and it is easier todistinguish the levels of colour from white to blue thanfrorn yellow to red-brown. The tetra-base method givesmuch better correlation with a quantitative method(r = 0.77) than does the picrate leaf test (r = 0.34),although the low value of the atter is partly due to poorcorrelation between leaf cyanide and tuber cyanide inthe same plant. The tetra-base method uses only small

. "arnounts of tetra-base which is not expensive (US$40per 100 g) and inexpensive cupric acetate. At least 100tests can be made for one cent.

Acknowledgements

We thank Dr. M. Bokanga of IIT A for many useful discussions andMr. Paul Ferrar of the Australian Centre for InternationalAgricultural Research (ACIAR) for rnuch assistance in the develop-rnent of a collaborative ongoing program that involves financia Isupport from ACIAR, IIT A and the Australian InternationalDevelopment Assistance Bureau (AIDAB).

REFERENCES

Bradbury, J. H. and Egan, S. V. (1991). Improved methods ofanalysis for cyanide in cassava and screening for lowcyanide varieties in the Pacifico Proc. Ninth Symp. Intern.Soe. Trop. Root Crops, Ghana, October 1991, in press.

Bradbury, J. H. and Holloway, W. D. (1988). Chemistry ofTropical Root Crops: Significance for Nutrition and

\

Agriculture in the Pacifico Australian Centre for InternationalAgricultural Research, Canberra, Australia.

Bradbury, J. H., Egan, S. V. and Lynch, M. J. (1991). Analysis ofcyanide in cassava using acid hydrolysis of cyanogenic

\glucosides. J. Sei. Food Agric. 55, 277-290.

Brinker, A. M. and Seigler, D. S. (1989). Methods for thedetection and quantitative determination of cyanide in plantmateriais. Phytochem. BulI. 21, 24-31.

Cooke, R. D. (1978). An enzymatic assav+or the total cyanidecóntent 5)f cassava. J. Sei. Food Agric. 29; 345-352.

Cooke, R. D., Howland, A. K. and Hahn, S. K. (1978). Screeningcassava for low cyanide using an enzymatic assay. J. Expl.Agric. 14, 367-372.

de Bruijn, G. H. (1971). A study of the cyanogenic character ofcassava. Mededelingen Landbouwhogeschool (Wagen-ingen) 171 (13), 1-140.

Delange, F., Bourdoux, P., Colinet, E., Courtois, P., Hennart, P.,Lagasse, R., Mafuta, M., Seghers, P., Thilly, C., Vanderpas,J., Yunga, Y. and Ermans, A. M. (1983). Nutritional factorsinvolved in the goitrogenic action of cassava. In CassavaToxicity and Thyroid: Research and Public Health lssues,(Delange, F. and Ahluwalia, R., eds). pp. 17-26. InternationalDevelopment Research Centre, Ottawa, Canada._J Esquivil, T. F. and Maravalhas, N. (1973). Rapid field method forevaluating hydrocyanic toxicity of cassaya root tubers. J.Agric. Fooa Chem. 21, 321-322.

-

94 J. H. BRADBURY AND S. V.EGAN.; .

Feigl, F. (1954). Spot test I. /norganic Applications, pp. 258-259. Mkpong, O. E., Yan, H., Chism, G. and Sayre, R. T. (1990).Elsevier, New York. . Purification, characterization and localization of linamarase

Feigl, F. and Anger, V. (1966). Replacement of benzidine by in cassava. Plant Physiol. 93, 176-181.copper ethylacetoacetate and tetra base as spot-test reagent Montgomery, R. D. (1980) Cyanogens, In Toxic Constituents offor hydrogen cyanide and cyanogen. Ana/yst 91,282-284. Plant Foodstuffs, (Liener, I. E., ed.) pp. 143-160. Academic

Fukuba, H. and Mendoza, E. M. T. (1981). Determination of Press, New York.cyanide in cassava, In Bioresources Investigation on Osuntokun, B. O. (1981). Cassava diet, chronic cyanide intoxica-Production, Storage, Processing and Vegetation of Root tion and neuropathy in the Nigerian Africans. World Rev,Crops in the Tropics, (Uritani, I. and Reyes, E. D., eds.) pp. Nutrition Oiet 36,141-173.48-62. Nagoya University Cooperation, Nagoya, Japan. Rosling, H. (1987). Cassava Toxicity and Food Seeurity. Tryck 1

Guignard, M. L. (1906). Le haricot a acide cyanhydrique, Kontakt, Uppsala, Sweden.Phaseo/us lunatus L. Compt. Rend. Paris 142,545-553. Sadik, S., Okereke, O. U. and Hahn, S. K. (1974). Sereening for

Howlett, W. P., Brubaker, G. R., Mlingi, N. and Rosling, H. (1990). Aeyanogenesis in Cassava, International Institute forKonzo, an epidemic upper motor neuron disease studied in Tropical Ariculture, Tech. Buli. No. 4.Tanzania. Brain 113, 223-235. Solomonson, L. P. (1981). Cyanide as a metabolic inhibitor, In

Ikediobi, C. O., Onyia, G. O. C. and Eluwah, C. E. (1980). A rapid Cyanide in Biology, (Vennesland, B., Conn, E. E., Knowles:.and inexpensive enzymatic assay for total cyanide in cas- C. J., Westley, J. and Wissing, F. eds.) pp. 11-28. Academicsava and cassava products. Agric. Biol. Chem. 44, 2803- Press, London.2809. Tantisewie, B., Ruijgrok, H. W. L. and Hegnauer, R. (1969). Die

lzomkun-Etiobhio, B. O. and Ugochukwu, E. N. (1984). Verbreitung der Blausaure bei den Cormophyten 5.Comparison of an alkaline picrate and a pyridine- Mitteilung. Pharm. Weekblad 104,1341-1355.pyrazolone method for the determination of hydrogen cya- Vogel, A. I. (1989). Textbook of Praetical Organic Chemistry, 5th

) nide in cassava and its products. J. Sei. Food Agrie. 35, 1-4. J ed., p. 1214. Longmans, London.Kojima, M., Iwatsuki, N., Data, E. S., Villegas, C. D. V. and Uritani, Williams, H. J. and Edwards, T. G. (1980). Estimation of cyanide

I. (1983). Changes of cyanide content and linamarase ac- with alkaline picrate, J. Sei. Food Agric. 31,15-22.tivity in wounded cassava roots. Plant Physiol. 72,186-189.

j