974 Analyst, October, 1980, Vol. 105, **. 974-980 Determination of Cyanide in Animal Feeding Stuffs J. R. Harris, G. H. J. Merson, M. J. Hardy* and D. J. Curtis Department of Industry, Laboratory of the Government Chemist, Cornwall House, Stamford Street, London, SE1 9NQ A method for the determination of cyanide in feeding stuffs has been developed. Naturally occurring cyano-substituted glycosides are subjected to enzymatic hydrolysis, the liberated cyanide is isolated by aeration and determined either by a spectrophotometric method or by gas chromatography. Recoveries of cyanide added to feeding stuffs at concentrations of 10 and 20 mg kg-' were approximately 98%. The method is sensitive to as little as 1 mg bg-1 of cyanide. Keywords; Cyanide determination; animal feeding stuffs; enzymatic hydvolysis; spectrofihotometry; gas chromatography Cyanide in trace amounts is found in a large number of plants, mainly in the form of cyano- substituted glycosides.Relatively high concentrations are found in certain grasses, pulses, roots and fruit kernels. Glycosides that have been positively identified include amygdalin in fruit kernels, especially bitter almonds; dhurrin in sorghum and other grasses; and linamarin in pulses, linseed and cassava. The structures and reactions by which these glycosides release free cyanide have been described by Montgomery1 and Conm2 Of the plant products mentioned above, sorghum, linseed (after extraction of linseed oil) and cassava are important ingredients in animal feeding stuffs but their use is restricted by the presence of the cyano- substituted glycosides.The EEC Directive3 on the control of undesirable substances in feeding stuffs prescribes maximum concentrations of hydrogen cyanide for both straight and complete feeding stuffs, ranging from 350 mg k g l (as hydrogen cyanide )for linseed cake down to 10 mg kg-l for complete feeding stuffs for chicks. The EEC method4 prescribed for the determination of cyanide involves enzymatic hydro- lysis, steam distillation and silver nitrate (Volhard) titration. The method is not specific for hydrogen cyanide and its lack of sensitivity is such that for concentrations of 10 mg k g l , only 0.2 ml of 1 N silver nitrate solution is consumed. This paper describes a procedure that is free from interferences and its sensitivity permits as little as 1 mg kg-I to be determined.As Cooke5 pointed out, most methods for the determination of cyanide have three stages: (i) hydrolysis of the cyano-substituted glycoside; (ii) isolation of the cyanide, for example by steam distillation or aeration; (iii) determination of the cyanide. This author states that most of the difficulties arise in stages (i) and (ii), as there are good techniques for cyanide determination; we would agree with Cooke although in our experience stage (ii), the isolation of the cyanide, is the most troublesome. Although methods for the determination of cyanide in plant materials have been published by C ~ o k e , ~ Wood,6 Blaedel et aL7 and Winkler,* none of these is directly applicable to compound animal feeding stuffs.In the determination stage Barkg considers that, for trace amounts of cyanide, the most suitable method is the spectro- photometric procedure based on the Konig synthesis of a pyridint. dyestuff by the reaction between cyanogen bromide and a suitable aromatic amine. In a further paper, Bark and HigsonlO recommended 9-phenylenediamine as the most suitable coupling agent ; this com- pound is specified in the method for the determination of cyanide in water,ll and has been adopted in the proposed method. More recently, gas-chromatographic m e t h ~ d s l ~ - ' ~ for trace amounts of cyanide have been described. In these methods the cyanide ion is reacted with bromine to produce cyanogen bromide, which, after separation on the gas chromatographic column, is determined by the use of an electron-capture detector.'This procedure is included in the proposed method as an alternative to the spectrophotometric prlocedure. * Present address: Beecham Pharmaceuticals Research Division, Great Burgh, Epsom, Surrey. Crown Copyright.975 HARRIS, MERSON, HARDY AND CURTIS Experimental Reagents Orthophosphoric acid, 10% VjV and 0.1 M. Hydrochloric acid, d = 1.18 g ml-l and 0.5 M. Sodium hydroxide solutions, approximately 0.5 and 1 M. Phosphate buffer solution, pH 7.0. in water and dilute to 1 1 with water. hydroxide solution. Dissolve 27.22 g of potassium dihydrogen orthophosphate Adjust the pH of the solution to 7.0 with 5 M sodium Bromine water. Saturated solution. Arsenious acid solution. Dissolve 2 g of arsenic(II1) oxide in 100 ml of water by boiling under reflux.Pyridine solution. Mix 150 ml of pyridine with 100 ml of water and add 25 ml of hydro- chloric acid (d = 1.18 g ml-l). p-Phenylenedaamine dihydrochloride solution. Dissolve 0.17 g of p-phenylenediamine hydrochloride (or 0.1 g of 9-phenylenediamine) in 50 ml of 0.5 M hydrochloric acid. p-Phenylenediamine - pyridine reagent. For each determination mix 15 ml of pyridine solution with 5 ml of p-phenylenediamine dihydrochloride solution immediately before the test. Diisopropyl ether. Phenol, 57; mjV aqueous solution. Potassium cyanide stock solution. Dissolve 1.563 g of potassium cyanide in water, add 25 ml of approximately 1 M sodium hydroxide solution and dilute to 250 ml. Standardise the solution against 0.1 M silver nitrate solution (1.0 ml of 0.1 M silver nitrate solution = 2.5 mg To 10.0 ml of stock solution add 50 ml of approximately 1 M sodium hydroxide solution and dilute to 500 ml (1.0 ml = 50 pg of CN-).To 10.0 ml of intermediate stock solution add 50 ml of approximately 1 M sodium hydroxide solution and dilute to 500 ml. Prepare the inter- mediate and working solutions freshly as required (1.0 ml = 1.0 pg CN-). Chop very finely or grind (preferably in a coffee grinder) 10 blanched sweet almonds and suspend in 100 ml of water. of CN-). Potassiuwz cyanide intermediate stock solution. Potassium cyanide aorking solution. Almond suspension. Shake the mixture thoroughly before use. Apparatus Aeration apparatus. The apparatus consists of a train of four flasks, each fitted with a Drechsel-bottle head.The reaction flask is of 500-ml capacity, round bottomed with a long neck. Preceding the reaction flask is a conical flask (250 or 500 ml) containing 0.5 M sodium hydroxide solution (to remove trace amounts of sulphur dioxide from the incoming air). Following the reaction vessel are two conical flasks (250 ml), each containing 50 mi of 0.5 M sodium hydroxide solution for absorption of the liberated hydrogen cyanide. (A Cecil CE595 double-beam ultraviolet spectrophotometer was used.) (A Pye Unicam GCV fitted with a nickel-63 electron-capture detector was used.) Operat- ing temperatures: column, 100 "C; injector, 125 "C; detector, 150 "C. The flask is immersed in a water-bath a t 37-40 "C. Spectrophotometer. Gas chromatograph. Column. With 10-mm cells, operated a t 515 nm.Fitted with an electron-capture detector. Silanised glass, 1 m x 4 mm i.d., containing Porapak Q (80-100 mesh). Procedure Extraction Weigh, to the nearest 0.005 g, approximately 20 g of the finely divided and mixed sample (previously ground to pass a 1-mm sieve) and transfer it into a 500-ml conical flask. Add by pipette 200 ml of 0.1 M orthophosphoric acid, stopper the flask and shake for 1 h. Allow the contents of the flask to settle and decant the supernatant liquid into a 250-ml centrifuge bottle and spin a t 2500 rev min-1 for 5 min. Transfer the supernatant liquid into a clean dry flask.976 HARRIS et al. : DETERMINATION OF Analyst, Vol. 105 Enzymatic hydrolysis Place approximately 250 ml of the phosphate buffer solution into the enzymolysis flask and add 10 ml of sweet almond suspension and two drops of antifoam.Introduce by pipette 50 ml of the sample extract prepared as described in the previouis section. Fit the bubbler system on to the enzymolysis flask and aerate the contents with an air flow of approximately 1.5 1 min-1 for 18 h, maintaining the temperature of the reaction mixture a t 3740 “C by means of a water-bath. Collect the liberated hydrogen cyanide in two 250-ml conical flasks arranged in series and each containing 50 ml of approximately 0.5 M sodium hydroxide solution. Quantitatively combine the solutions in a 200-ml calibrated flask, dilute to the mark with water and mix. Blank of the sample extract and following the procedure described under Enzymatic hydrolysis.Determination Spectrophotometric method Into a 50-ml calibrated flask transfer an aliquot portion of the sample solution obtained from the enzymatic hydrolysis, not exceeding 20 ml and expected to contain not more than 20 pg of cyanide (as CN-). In a second 50-ml calibrated flask place an identical volume of the hydrolysis blank solution. To each flask add 0.5 ml of hydrochloric acid (d = 1.l8gm1k1), mix and then immediately add 0.5 ml of bromine water. Stopper the flasks to prevent losses of cyanogen bromide, mix and allow to stand for 2 min. Add to each flask 0.5 ml of arsenious acid, mix thoroughly to remove excess of bromine, then add 20.0 rnl of p-phenylenediamine - pyridine reagent, dilute to the mark with water, mix and allow to stand for 40 min. Without further delay measure the absorbance of the solutions at 515 nm in 10-mm cells against water as reference.Subtract the absorbance of the hydrolysis blank from that of the sample and determine the amount of cyanide in the sample solution by reference to the calibration graph. Transfer by pipette into a series of 50-ml calibrated flasks 5, 10, 15 and 20 ml of the cyanide working solution and dilute as necessary to 20 ml with water. In addition include one flask containing 20 ml of water without cyanide as the reagent blank. Proceed as described above from “To each flask add 0.5 ml of hydrochloric acid . . .” Measure the absorbance of each solution after 40 min a t 515 nm in 10-mm cells against water as reference. Subtract the absorbance of the reagent blank from those of the cyanide standards and plot the calibration graph using the net absorbance values on the ordinate atnd the corresponding masses of cyanide (as micrograms of CN-) on the abscissa.Gas-chromatographic method Into a 100-ml separating funnel transfer an aliquot portion of the sample solution obtained from the enzymatic hydrolysis, not exceeding 20 ml and expected to contain not more than 8 pg of cyanide (as CN-)*. Into a second separating funnel place an identical volume of the hydrolysis blank solution. Make up the volumes to 20 ml if necessary. Add 5 ml of 10% VjV orthophosphoric acid solution and 0.5 ml of bromine water. Shake each funnel and allow to stand for 15 min, then add 0.2 ml of 5% m/V aqueous phenlol solution and gently shake. Add 20.0 ml of diisopropyl ether to each funnel, stopper firmly and shake vigorously for 2 min.Allow the phases to separate, then discard the lower aqueous phase and transfer the diisopropyl ether layer into a vial fitted with a septum cap. Inject 1.0 pl of each solution on to the column and measure the heights of the cyanogen bromide peaks obtained. Subtract the peak height of the cyanogen bromide in the hydrolysis blank solution (if any) from that of the sample and determine the amount of cyanide in the solution by reference to the calibra- tion graph. Calibration graph. Transfer by pipette into a series of 100-ml separating funnels 2, 4, 6, 8 and 10 ml of cyanide working solution, dilute as necessary to 20 ml with water. In addition * The amount of cyanide in the solution injected into the gas chromatograph must be within the linear range of the detector.In some instances it may be necessary to reduce both the recommended maximum content in the sample solution and the amounts prescribed for the calibration graph in order to meet this requirement. Carry out a blank enzymatic hydrolysis using 50 ml of 0.1 M orthophosphoric acid in place Calibration graph.October, 1980 CYANIDE IN ANIMAL FEEDING STUFFS 977 include one separating funnel containing 20 ml of water without cyanide as the reagent blank. Proceed as described above from “Add 5 ml of 10% VjV orthophosphoric acid. . .” Measure the heights of the cyanogen bromide peaks and, after correcting for the reagent blank, plot the calibration graph of corrected peak height against the corresponding mass of cyanide (as micrograms of CN-).Calculation The same calculation is applicable to both procedures. The cyanide content of the feed in milligrams per kilogram is given by 800 x A M x V where A = mass of cyanide (as micrograms of CN-) present in the aliquot portion of the sample extract, M = mass of test portion (grams) and V = volume of the aliquot portion taken for the determination (millilitres) . Results and Discussion Recovery of Cyanide from Cyano-substituted Glycosides The efficiency of the proposed method was assessed by adding to the orthophosphoric acid used in the extraction stage known amounts of amygdalin (Aldrich Chemical Co. Ltd; cyanide content 5.69% m/m as CN-) and linamarin (Calbiochem, Bishops Stortford, Hertfordshire; cyanide content 10.52% mlrn as CN-) and following the procedure described above.The results are given in Table I and show that near-quantitative yields of cyanide were obtained. TABLE I RECOVERY OF CYANIDE FROM AMYGDALIN AND LINAMARIN IN THE ABSENCE OF FEED Spectrophotometric method GC method 1 CK- added r------- - r---L--- Glycoside PLg P8 % Pg % as substratel CN- found/ Recovery, CN- found/ Recovery, Amygdalin. . 200 200 197 201 400 405 389 397 Linamarin . . 200 194 197 199 191 191 198 100.0 98.5 100.5 101.3 97.3 99.3 97.0 98.5 99.5 95.5 95.5 99.0 197 98.5 199 99.5 196 98.0 - - - - 196 98.0 201 100.5 197 98.5 - - Having established the efficiencv and reliability of the method when applied to amygdalin and linamarin, it was then applied to a series of compound feeding stuffs fortified with these two glycosides.The results of these studies are shown in Table 11, from which the recovery of cyanide and the repeatability of the method can be seen to be good. The mean recovery by the spectrophotometric method was 97.9% with a standard deviation of 1.24%. Corres- ponding figures for the gas-chromatographic method were 98.2% and 0.98%, respectively. For samples where both the spectrophotometric and gas-chromatographic methods were used a comparison of the results showed no significant difference between the procedures. No cyanide was detected in the unfortified feeds. Determination of Cyanide in Straight and Compound Feeding Stuffs Several straight feeding stuffs and a compound feeding stuff containing linseed expeller were examined by the proposed method.The results shown in Table I11 demonstrate the good repeatability of the method and the high level of agreement between the spectrophoto- metric and gas-chromatographic procedures.978 HARRIS et al. : DETERMINATION OF Anallyst, Vol, 105 As a further check on the recoveries of cyanide the method was applied to the products listed in Table I11 after fortification with amygdalin and linamarin ar concentrations equivalent to 10 mg k g l of CN-. Before these experiments could be conducted it was necessary to remove the cyanide arising from natural glycosides present. This was achieved by subjecting the products to the extraction and enzymatic hydrolysis procedures described above before adding known amounts of amygdalin and linamarin and then performing a second enzymolysis.Unfortified extracts treated similarly yielded no cyanide, indicating that the first enzymolysis was effective in removing all the cyanide initially present. The results of these experiments are given in Table IV and demonstrate the good recoveries given by the method. TABLE I1 RECOVERY OF CYANIDE FROM ANIMAL FEEDS FORTIFIED WITH AMYGDALIN AND LINAMARIN Feed (20.0-g sample) Glycoside Chick mash A . . Amygdalin Chick mash B Linamarin Linamarin Layers mash. . . . Amygdalin Linamarin Dairy ration containing approximately 18% m/m of rape meal Amygadlin Linamarin Spectrophotometric method GC method 7 r---.LL-- , ---A CN- added/ CN- found/ Recovery, CN- found/ Recovery, Pg Pg % Pg % 200 196 98.0 194 97.0 197 98.5 199 99.5 197 98.5 197 98.5 191 95.5 194 97.0 197 98.5 194 97.0 ~.~ 191 95.5 200 194 97.0 196 98.0 196 98.0 193 96.5 400 395 98.8 400 100.G 386 96.5 38 1 95.3 386 96.5 391 97.8 200 195 97.5 199 99.5 - _ _ 197 98.5 200 200 100.0 197 98.5 197 98.5 200 195 97.5 197 98.5 197 98.5 200 197 98.5 195 97.5 198 99.0 199 99.5 195 195 198 97.5 97.5 99.0 Liberation and Isolation of Cyanide In preliminary investigations the enzymatic hydrolysis was allowed to proceed for about 16 h at 37 "C in a sealed flask. At the end of this period the liberated cyanide was steam distilled from the reaction mixture under weakly acidic conditions. Although quantitative yields of cyanide from amygdalin and potassium cyanide were recorded in the absence of a feed matrix, the inclusion of such a matrix (ground wheat) reduced the recovery to 80-90% of the theoretical yield of cyanide.When potassium cyanide was added to ground wheat and the mixture was distilled immediately complete recovery of cyanide was obtained. These observations indicated the possibility of a reaction between the liberated cyanide and the feed matrix. If the cyanide could be removed from the reaction mixture whilst the enzymatic hydrolysis was in progress then improved yields might be obtained. An aeration technique was therefore employed with an air flow-rate of 1.5 1 min-l and the liberated cyanide collectedOctober, 1980 CYANIDE I N AXIMAL FEEDING STUFFS 979 in 0.5 11 sodium hydroxide solution. Improved yields were obtained but these were not consistent and on re-checking the yield of cyanide from amygdalin in the absence of a feed matrix a result of about IIOo/o was recorded using the spectrophotometric procedure.Further investigation showed that as a result of using an acetate buffer to maintain a pH of 5.0 (as recommended by the suppliers of the P-glucosidase enzyme used a t this stage of the work) TABLE 111 RESULTS OF ANALYSIS: CYANIDE CONTENT OF FEEDS AND NATURAL PRODUCTS CK- found/yg g-l r-------L------ 7 Spectrophotometric Matrix method GC method Expeller linseed . . . . 380, 392, 395, 392 385, 380, 378, 370 Feed containing 4% mjm linseed . . . . . . 8.6, 8.5, 8.3 8.6, 8.6, 8.9 Ground cassava tubers . . 159, 157, 150 - Sorghum (seeds) . . . . 1.5, 1.4, 1.5 - -4lfalfa . . . . . . . . 4.3, 4.0, 4.1 - acetic acid was being carried over during the aeration stage and was enhancing the absorbance readings.This effect could not be compensated for by making blank determinations as it only became manifest in the presence of cyanide. When the acetate buffer was replaced by a phthalate buffer at pH 5.0 the carry-over problem was eliminated. In applying the method to expeller linseed using phthalate buffers a t varying pH values, the maximum yield of cyanide was obtained at pH 7.0. Following the work of C ~ o k e , ~ the phthalate buffer was sub- sequently replaced by one containing potassium dihydrogen phosphate at pH 7.0. The method described by Cooke5 employs extraction of the cyano-substituted glycosides by orthophosphoric acid, thereby obviating the need for steam-distillation or aeration. Cooke's method did not appear to be wholly applicable t o animal feeding stuffs but the principle of extracting the glycosides from the feed before proceeding with the enzymatic hydrolysis aeration stage offered a further means of reducing the matrix effects discussed above.The orthophosphoric acid extraction stage was therefore combined with the aeration procedure and gave consistent, quantitative yields, as shown in the tables. TABLE IV RECOVERY OF CYANIDE FROM FEED MATERIALS FORTIFIED WITH AMYGDALIN AND LINAMARIN Matrix Glycoside Expeller linseed . . , . Amygdalin Linamarin Feed with 4% mjm linseed . . Amygdalin Linamarin Cassava . . . . . . Amygdalin Linamarin Sorghum (seeds) . . . . Amygdalin Linamarin Alfalfa . . . . . . . . Amygdalin Linamarin Recovery, yo r---------h------- 7 Spectrophotometric 95.7, 98.7, 97.4 96.4, 95.9, 97.3 98.0, 98.5, 99.6 94.2, 97.6, 97.4 96.3, 96.8, 99.1 - 95.9, 96.3, 97.1 - 97.8, 96.3, 96.9 - 95.6, 96.5, 95.0 - 92.3, 94.3, 95.1 - 93.9, 95.6, 96.2 - method GC method 96.5, 98.3, 97.1 97.1, 97.5, 96.9 97.3, 97.3, 96.4 95.9, 95.1, 96.7 Enzyme Sources The EEC method4 for cyanide in feeding stuffs prescribes the use of blanched sweet almonds as the enzyme source in the liberation of cyanide from the glycosides.As purified enzymes are now readily available it was decided a t the start of this work to use commercial P-D-glucosidase (E.C. 3.2.1.21) in place of the sweet almond suspension with a view to standard- ising the reagent. Although cyanide was liberated from expeller linseed in the presence of purified P-glucosidase, little or none was obtained from linamarin, the cyano-substituted glycoside of linseed.However, by using as the enzyme source an aqueous suspension of sweet almonds, good yields of cyanide were obtained from linamarin. The reasons for the980 HARRIS, MERSON, HARDY AND CURTIS very low yields of cyanide when p-D-glucosidase was used are not known, one explanation might be that in preparing the pure enzyme (from almonds) trace amounts of other materials, including enzymes necessary for the hydrolysis of linamarin, have been removed. The fact that some cyanide was liberated from expeller linseed in the presence of purified P-D-ghcosidase was undoubtedly due to the presence of endogenous enzymes, the added enzyme being superfluous to the reaction.From these observations it was concluded that in a general method. for the determination of cyanide in feeding stuffs, an enzyme system capable of acting on several glycosides must be present. The sweet almond suspension seems to supply the necessary range of enzymes and has therefore been retained in the proposed method in preference to pure enzyme sources. We thank the Government Chemist for permission to publish this paper. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. References Montgomery, R. D., in Liener, I. E., Editor, “Toxic Constituents of Plant Foodstuffs,” Academic Conn, E. E., in “Toxicants Occurring Naturally in Foods,” National Academy of Sciences Publication Council Directive 74/63 EEC, Oficial Jour.rta1 ofthe European Communities, L38, 1 l t h February 1974, Commission Directive 71/250 EEC, Oficial Journal of the European Communities, L155, 12th July Cooke, R. D., J . Sci. Food Agric., 1978, 29, 345. Wood, T., J . Sci. Food Agric., 1965, 16, 300. Blaedel, W. J., Easty, D. B., Anderson, L . , and Farrell, T. R., Anal. Chem., 1971, 43, 890. Winkler, W. O., J . Assoc. Off. Agric. Chem.. 1951, 34, 541. Bark, L. S., I n d . Chem., 1962 (October), 255. Bark, L. S., and Higson, H. G., Talanta, 1964, 11, 471. Department of the Environment, “Analysis of Raw, Potable and W%ste Waters,” HM Stationery Nota, G., and Palombari, R., J . Chromatogr., 1973, 84, 37. Nota, G., Palombari, R., and Improta, C., J . Chromatogr., 1976, 123, 411. Bates, B. L., and Buick, D. R., J . Assoc. Off. Anal. Chem., 1976, 59, 1390. Press, New York, 1969, p. 143. 1354, Natural Research Council, Washington, D.C , 1966. p. 31. 1971, p. 13. Office, London, 1972, p. 227. Received March 31st, 1980 Accepted May 28th, 1980