Biochemtil Systematics and Ecology, Vol. 14, No. 1. pp. 97-103.1986. Printed in Great Britain.

0305-1978/86 $3.00+0.00 Pergamon Press Ltd.

Simple Densitometric Method for Estimation of Cyanogenic Glycosides and Cyanohydrins Under Field Conditions

LEON BRIMER and PER MBLGAARD

Department of Pharmacognosy and Botany, Royal Danish School of Pharmacy, 2 Universitetsparken, DK-2100 Copenhagen 0, Denmark

Key Word Index-Cyanogenic glycosides; cyanohydrins: detection; quantitative determination; field method; Thalictrum kwdleri.

Abstract-In this new design of the picrate test (Guignard Method) for small scale cyanide determination under field conditions, hydrolysis of the cyanogenic compounds is catalysed by ‘@glucuronidase’, a commercial enzyme preparation from Helixpomatia. The method described facilitates the quantification of small amounts of cyanogenic compounds in plant material, with a detection limit about 60 ng of released HCN (equivalent to 1 pg of amygdalin). False positive tests produced by hydrolysis products of gluco- sinolates could be excluded by inhibition of the myrosinae enzyme system present in plants that contain glucosinolates. The total weight of equipment needed for a total of loo0 determinations (in groups of eight) does not exceed 500 g, thus making the method convenient for field work.

Introduction Only a few reports of the design of methods for phytochemical screening in the field are found in recent literature. Usually, the methods employed are only qualitative, or, at best, semiquantitative, and they often involve a visual comparison with standards and the results are expressed as + to ++++ [I], or unusually, by a scale O-9 [2].

Cyanogenic constituents (i.e. cyanogenic glycosides, cyanogenic lipids and cyanohydrins) are often included in such screening programs. Detection of cyanogenesis, as well as the quali- and quantitative estimation of cyanogenic com- pounds in plants and insects, is traditionally based on reactions specific for HCN released after cleavage of the cyanogenic constituents [3]. However, enzyme preparations able to hydrolyse all types of cyanogenic glycosides have not been readily available [4,51 and quantitative release of HCN from the matrix of organic compounds may be incomplete [31. Until now, nearly all known methods for quantitative determination of HCN have been inconvenient for use in the field. Thus, Kaplan et a/. [6] recently stated the need for a more reliable field test system.

As a possible way to solve these difficulties, we explored the combination of a modified

(Received 25 January 1985)

microdiffusion based picrate (Guignard) test, with the use of ‘j3-glucuronidase’ from Helixpom- atia as a source of hydrolytic enzymes [5]. In the modified picrate test, released HCN is detected as red to brown spots on transparent hydro- phobic picrate reagent sheets (Fig. 1). This sys- tem permits densitometric recording of colour intensities corresponding to quantities of HCN as low as 60 ng released from 50 mg, or less of

d

FIG. 1. SKETCH OF MICRODIFFUSION CHAMBER FOR DETERMINATION

OF RELEASED CYANIDE. (a) Material ground in enzyme solution

together with a drop of CHCI, (to promote cell lysis); (bl Transparent re-

agent sheet cover; (c) Picrate impregnated layer of the reagent sheet;

(di Part of reagent sheet after exposure to HCN, showing an orange-red

to violet welldefined spot. to be rated visually or measured with a

densitometer.

97

LEON BRIMER AND PER M0LGAARD

plant material. This corresponds to about 1.2 pg HCN/g. From a chemotaxonomic point of view, plants are normally considered cyanogenic if they release more than IO-20 frg of HCN/g fresh material [7].

Results and Discussion Two different designs (1 and 2, see Experi- mental) were developed for qualitative and quantitative determination of liberated HCN, respectively. The two designs are satisfactory for a set of basic demands for accuracy and repro- ducability and still the equipment is easy to use for field trials. A large number of samples can be screened simultaneously with a detection limit as low as 60 ng of liberated HCN, permitting use of small samples (50 mg or less) and study of weakly cyanogenic specimens. For field tests it is important that part of the equipment is re- usable, sturdy and easy to clean, although of light-weight material. The advantage of the quantitative method is the dual rating of the re- agent strips, in the laboratory with a densitome- ter and in the field by comparison with a set of standards on reagent sheets.

The densitrometrically determined continuous calibration curve (Fig. 2, Design 2) was obtained by means of standard solutions of amygdalin, and compared with the curve obtained, when

15r

I

. 8 . : .

i 1; , , ,:, 2bo

0 50 loo

Amount of substrate (n ml)

FIG. 2. CALIBRATION GRAPH FOR MICRODIFFUSION OF INCREASING

AMOUNTS OF RELEASED CYANIDE-CLOSED DOTS. The developed

picrate sheets have been read with a densitometer. Enzyme concentra-

tion 2% v/v, 24 h, 21”. The dots represent means of three replicates +

standard deviation. Open dots present results from a single experiment

with standard solutions applied to silica gel plates, which were sprayed

with a 6% vlv enzyme solution and covered with the reaction sheet.

I5 r

Amount of substrata (n mol)

FIG. 3. EFFECT ON THE CALIBRATION GRAPH OF THE PRESENCE OF

PLANT MATERIAL. Conditions as in Fig. 2 (filled dots) and with addition

of 50 mg of fresh leaves of Geranium grevekanum (squares) and

Plantego major (open dots). The curve represents the calibration graph

from Fig. 2.

solutions were applied to silica gel plates, sprayed with a 6% v/v enzyme solution and covered directly with the reaction sheet. This later method has previously been shown to cause total hydrolyses under these conditions [5] and since the two calibration curves are identical, the reaction was considered quantitative in the case of the microdiffusion model also. However, experiments including addition of different kinds of non-cyanogenic plant material clearly proved that the matrix could impede the release of HCN, as previously described [3]. Species of Geranium proved to be very efficient in this respect, and G. grevelleanum Wall. (among other species) was used for the investigation of incubation parame- ters, in order to establish a set of standard condi- tions which permit the quantitative determina- tion of most cyanogenic plant material (Fig. 3, Table 1).

The assay was checked both qualitatively (Table 2) on different plant material (low-, high- and acyanogenic) and quantitatively on leaves from Prunus laumcerasus L. cv. schriphaensis. Quantitative HCN-determination with Prunus leaves was performed according to the estab- lished ‘Recommended standard assay (2)’ (Experimental) with 50 mg of fresh material and compared with a traditional calorimetric deter- mination of HCN from 200 mg of fresh material collected by aeration in NaOH (2 ml, 0.1 M) [81. The results from four replicates of the densito- metric and the calorimetric determination of

ESTIMATION OF CYANOGENIC COMPOUNDS 99

TABLE 1. COLOUR RESPONSE TO ENZYME CONCENTRATION, TIME

AND TEMPERAlURE IN STANDARD INCUBATION OF AMYGDALIN (65

“mol)*

Enzyme Time of Colour response

concentration exposure Temperature equivalent to

(% vlv/F.u.ml ‘) (hl (‘C) released HCN (nmol)

2/2oQo 16 21 12fl

2/2goo 40 21 21f3

4/4000 16 21 25k3

2/2000 16 40 51f4

‘65 nmol of amygladi” with addition of Geranium grevelleanum leaves (50 mg) as HCN retaining material. Results of four replicates *standard

deviation.

liberated cyanide were 76f4 f.tg HCN/g and 77f 5 pg HCN/g, respectively.

However, due to a possible retaining of liber- ated compounds in the complex matrix of plant material and enzyme solution in the reagent jars, a proper quantitative analysis must comprise at least two independent but identical determina- tions of HCN under different conditions in respect of enzyme concentration, temperature or time.

Authentic sinignn as well as selected plant material known to contain glucosinolates that yield steam volatile products of hydrolysis, were investigated by this standard assay. Degradation products from these compounds are reported also to form orange red products with picrate (Table 2). All species previously reported to be cyanogenic gave the expected fast colour reac- tion, while sinigrin (with addition of myrosinase) and some of the species known to contain only glucosinolates showed a slow but definite reac- tion. In order to eliminate false positive tests, a qualitative confirmatory test was developed, based on extractive acid denaturation of any myrosinase that may be present in the plant material, prior to incubation. This procedure is a modification of a method used for extractions and determination of cyanogenic compounds in cassava roots [9]. Results for selected species proved this test valuable (Table 2), and warranted the rejection of the noncyanogenic species found positive in the standard assay, while all cyanogenic species remained positive.

Thalicbum fkndleri Engelm., which has not pre- viously been reported cyanogenic, appears to contain at least two cyanogenic compounds, when examined by thin-layer chromatography.

Finally, two standard strips were produced by placing a piece of silanized filter paper (J. C. Binzer, Hatzfeld/Eder, GFR) between the picrate reaction sheets and the reaction chamber. In one of these experiments, the rack was placed upside-down, which without the silanized filter paper would have completely soaked the reac- tion sheet. Both of these strips were found nearly identical with strips obtained by the stan- dard procedure when read densitometrically. Thus, a protective layer of this type may be introduced, in order to protect the reaction sheet, when transported during development.

The methods described seem to be well suited for determination of total content of cyanogenic glycosides and/or cyanohydrins in plant material under field conditions as well as in the laboratory. When rated visually, experiments in our laboratory showed that all the standards below 150 nmol used for the calibration graph shown in Fig. 2 could be arranged in the proper order by five individuals aged 2B-54. However, one disadvantage should be mentioned, namely the relatively short dynamic range and the result- ing difficulties encountered with strongly cyano- genie materials, since very small samples must be used in order to obtain quantitative results, i.e. to stay below 100 nmol of released HCN.

Experimental &sign 1 IF&. 4). Consists of a number of screw cap vials pro- vided with the corresponding open-top caps and placed in a light-weight rack. Circular discs of reagent sheet are made by means of a hollow punch or cork borer and placed into caps from which the septa have been removed. Colour reactions may be observed during exposure, due to the transparency of the sheet, and developed discs can be stored easily in small tubes for subsequent comparison with standards.

Design 2 (Fig. 5). Consists of microtest tubes (35 X 12.5 mm, polypropylene lymphocyte tubes) placed in a block so that the open ends of the tubes are level with the plane of the upper surface of the block, allowing a rectangular strip of reagent sheet to be placed as a close cover on top of the tubes. The sheet is pressed against the surface by means of a transparent cover (e.g. Perspex”). The spots on the developed reagent strip may be rated in the field and/or measured with a den- sitometer after return to the laboratory. Developed strips should be protected from light and corrosive vapours in a closed container or mounted between two glass plates. A standard strip for comparison in the field may be preserved in an envelope of transparent heat sealed plastic and kept pro- tected from light.

Preparation of reagent sheets. Precoated ion-exchange sheets (Polygram lonex 25SE-AC, Macherey-Nagel, D&en, Germany) were impregnated by consecutive immersion in

100 LEON BRIMER AND PER M O L G A A R D

TABLE 2, QUALITATIVE HCN DETERMINATION, BY MEANS OF 'RECOMMENDED STANDARD ASSAY' AND 'CONFIRMATORY TEST'

Species and origin* Part Amount Colour reaction:l:

tested1 (mg) after (h) cyanogeniet

24 48 120 compounds

Constituents as compi led from literature

glucosinolatest

Confirmatory

test ( / + colour

reaction) 24 h

Reference

Barbarea intermedia Bor. (4851/6)

Barbarea rntermedia Bor. (4851/6)

Brassica ]uncea (L.) Zern. et Coss.

(commercial)

Brassicaluncea (L) Zern. et Coss.

(commercial)

Brassicajuncea (L.) Zern. et Coss.

(commercial)

Brasslca nigra (L.) Koch. (4949/19)

Brasstca nigra (L.) Koch. (4949/19)

Dl#lotaxls tenuifo#a {L.) D.C. (4955/5)

Diplotaxts tenuifolia (L.) D.C. (4955/5)

Geramum grevelleanum Wall. (S 1971 /

1126)

Isatls tmctoria L. (4938/4)

Isatis tinctona L. (4938/4)

IsatJs tmctona L. (4938/4)

/saris tinctoria L, (4938/4)

L iriodendron tulipifera L. (4740/1)

Lunana redJvlva L. (4863)

Lunana redlviva L. (4863)

Lunaria rediviva L. (4863)

Lunaria rediviva L. (4863)

Malus baccata Borkh. var. oblonga

(6342F/2)

Malus domestlca Borkh. (commercial)

Nandina domestma Thunb. ($1971/

0577)

Passiflora quandrangularis L. (5098/19)

Ptantago major L (2170/182)

Prunus lauraceracus L cv.

Schripkaensis (P1974/5112)

Sambucus ebulus L. (RDSP)

Sambucus mgra L. (own)

Taxus baccata L. (~ (18168/1)

Thahctrum aquilegifolium L. (4772/3)

Thalictrum fendleriEngelm, (4772/21)II

Thahctrum foetidum L. (4772/7)

7halictrum glaucum Desf. (4772/8)

Thalictrum polygamum MQhl. (4772/20)

I 50 0 0

I 200 0 0

s 10 0 0

s 20 0 0/ I

100 2 3

50 0 O

200 0 0/1

50 0 0

200 0 0

50 0 0

50 0 0

200 0 0/1

50 0 0

200 0 0/1

50 4

0

0 gluconasturt i in §,s

0 sinigrin§,l 10

3

0

0 - - sinigrin§,l. 10

0

0 glucoerucin§,l. 10

0 - -

0/1

0 gluconapin§,s

0 4 , s h +,1

2

tr iglochinin

taxiphyl l in, i

50 0 0 0

200 0 0 0

50 0 0 0

200 0 0 0/1

glucoberteroin§,s tO

50 0 0 0 + , s -- 11

50 ;>5 4, s - 11

50 4/5 nandinaglucosid, I 1!

50 2 f , t/s. - - ~ 11

50 0 0 0 --

50 > 5 prunasin, I 11

50 0 0 0 - -

50 5 sambunigrin, prunasin, t 11

zierin, holocalin, I.

50 4 taxiphyl l in, t. -o + 12

50 5 t r ig lochininmono

methylester, proteacin,

nandinaglycosid, I. - - 11

50 3 3 3 --

50 0 0 0 ~, I - - 11

50 0 0 0

50 1 tr iglochinin, I 11

t0, 11

11

*The samples were either: collected August 3rd in Botanical Garden, University of Copenhagen, parenthesis indicates plant no , obtained through normal commerce (commercial), collected in the garden of Royal Danish School of Pharmacy (RDSP}, or in the author's own garden {own)

t l , leaves; s, seeds; fr, fruit; gl, green leaves; sh, shoot.

:l:Colour reaction rated visually against five standards equivalent to 1 ,= 60, 2 - 180, 3 600, 4 1800, 5 -- 6000 ng of HCN.

§Steam volatile products of hydrolyses.

II Not previously reported cyanogenic.

101

FIG. 4. MICRODIFFUSION SYSTEM, DESIGN 1. Light weight rack for 15 vials in a plastic box. An open-top cap and a vial ere seen to the right (arrow). A number of exposed discs may be seen on top of vials in the rack. In front, a number of discs have been exposed to HCN from differ- ent amygdalin standards.

102

FIG. 5. MICRODIFFUSION SYSTEM, DESIGN 2. Rack made of PVC. Re- agent strip (arrow) exposed to different amygdalin standards. Stabilizing aluminium-profile cover is seen to the right of the perspex cover.

ESTIMATION OF CYANOGENIC COMPOUNDS 103

three solutions: (1) a saturated soln of picric acid in H20, fol- lowed by air drying; (2) a 1 M aq. Na2CO 3 soln, followed by air drying; and (3) a 2% w/v ethanolic 1-hexadecanol soln, fol- lowed by air drying [5]. Calibration graphs should be produced for each batch of impregnated sheets due to small differences among batches.

Assays. A suitable amount of material (based on preliminary tests) was placed in a vial (Design 1) or tube (Design 2). One drop of CHCI3--to promote cell lysis--and 500 p.I of an aq. soln of I~-glucuronidase from Helix pomaga (Sigma G-0876) were added. The material was ground in the liquid with a glass pestle to form a homogeneous mixture. Finally the reaction chamber was closed by means of the reagent sheet, and left until complete reaction had occurred, in general 24 hours. Absorbance of the spots may be rated visually by comparison with standards or measured by the transmission technique with a densitometer [Vitatron TLD 100; light source: tungsten; secondary filter; 540 nm; mode: --log; aperture: 2.0 mm (circular)].

Spectrophotometric determination of total HCN released. HCN, released from material ground in a 2% aq. soln of I~- glucuronidase Helix pomatia was collected in 2 ml of NaOH by aeration with N 2 [3]. Total cyanide released after 24 h at 35 ° was determined according to Epstein as modified by J~rgensen [8].

Confirmatory test. Plant material (500 mg) was ground in a mortar with orthophosphoric acid (0.05 M, 2.5 ml). The homogenate was filtered by pressure through silica gel 60 (400 mg) placed on top of glass wool in a 2 ml syringe. Phosphate buffer (0.2 M, pH 7.0, 1.0 ml) and a 20% v/v aq. soln (50 p.I) of I}-glucuronidase Helix pomatia were added to the filtrate (0.5- 1.0) in a vial/tube. The resulting pH was about 6.0-6.8. The reaction chamber was closed by means of the reagent sheet.

Recommended standard assay. (1) Qualitative determination: amount of material, 50 mg; enzyme concentration, 2% v/v (corresponding to about 2000 Fishman units per ml); time and temperature, about 20 h for temperatures above 10-15 °.

(2) Quantitative determination--laboratory conditions: amount of material, decided after the colour response from a pre- liminary qualitative determination; enzyme concentration, 3-4% v/v (about 3-4000 Eu. per ml); time and temperature, 20 h at 40 °. (3) Quantitative determination--field conditions: amount of material, as above; enzyme concentration, 8% v/v (about 8000 F.u, per ml); time and temperature, 20 h at temperatures above 15-20%.

Acknowledgements--The authors thank Dr. S. Brogger Christensen and Dr. F. Nartey for critically reading the manuscript.

R e f e r e n c e s 1. Farnsworth, N. R. (1966) J. Pharm. Sci. 55, 225. 2. McGregor, D. I. and Downey, K. (1975) Can. J. Plant Sc~ 55,

191. 3. Nahrstedt, A., Erb, N. and Zinsmeister, H.-D. (1981) in Cyan-

ide in Biology (Vennesland, B., Conn, E. E., Knowles, C. J., Westley, J. and Wissing, F. eds) p. 461. Academic Press, London.

4. H6sel, W. (1981) in Cyanide in Biology (Vennesland, B., Conn, E. E., Knowles, C. J., Westley, J. and Wissing, F., eds) p. 217. Academic Press, London.

5. Brimer, L., Bragger Christensen, S., Molgaard, P. and Nartey, F. (1983) J. Agric. Food. Chem. 31, 789.

6. Kaplan, M. A. C., Figueiredo, M. R. and Gottlieb, O. R. (1983) Biochem. Syst. Ecol. 11, 367.

7. Hegnauer, R. (1977) Plant Syst. Evol. Suppl. 1, 141. 8. Jergensen, K. (1955) Acts Chem. Scand. 9, 548. 9. Cooke, R. D. (1978) J. Sci. FoodAgric. 29, 345.

10. Kje~r, A. (1960) Fortschr. Chem. Org. NatursL 18, 122. 11. Tjon Sie Fat, L. A. (1979) Contribution to the Knowledge of

Cyanogenesis in Angiosperms, p. 18. Dissertation, Univers- ity of Leiden, The Netherlands.

12. Hegnauer, R. (1959) Pharm. Weekbl. 94, 241.