A Straightforward Method for the Colorimetric Detection of

Oct 20, 2009 - ... for Instantaneous Monitoring of Cyanide Based on an Elsner-Like Reaction .... and Matthew Brenner , William Blackledge and Gerry R...
0 downloads 0 Views 266KB Size
Download PDF
Anal. Chem. 2009, 81, 9493–9498

A Straightforward Method for the Colorimetric Detection of Endogenous Biological Cyanide Christine Ma¨nnel-Croise´, Benjamin Probst, and Felix Zelder* Institute of Inorganic Chemistry, University of Zu¨rich, Winterthurerstrasse 190, 8057 Zu¨rich, Switzerland Corrin-based chemosensors allow the rapid and selective colorimetric detection of endogenous biological cyanide. The color change from orange to violet can be easily observed with the “naked eye” (∆λmax ) 51 nm). The methodology works directly in the biological matrix without time-consuming sample pretreatment and the use of special reaction conditions. It was possible to study the enzymatic release of cyanide from its biologicalprecursorlinamarinwithdiffusereflectanceUV-visible (DRUV-vis) spectroscopy on a freshly prepared biological surface. These experiments were accompanied by stopped-flow kinetic measurements under homogeneous conditions. Detection in the biological sample is based on the selective coordination of cyanide to the metal-based chemosensor as proven by UV-vis and 1H NMR spectroscopy. Examples of applications during food manufacturing are given. Hydrogen cyanide is highly toxic to humans and almost all other forms of life.1,2 On the other hand, it is widespread in nature, either produced and further metabolized by cyanogenic microorganisms or integrated and stored as cyanogenic glycosides in higher plants like sorghum, flax, giant taro, bamboo, and cassava.3,4 The latter one is one of the most important carbohydrate sources for about 500 million people in South America and parts of Africa.5,6 The enzymatic release of cyanide from cyanogenic glycosides after cell rupture (Scheme 1) exhibits great danger to the consumer and can cause severe chronic as well as significant acute public health problems.7,8 Therefore, the detection and removal of biological cyanide is of general importance.6,9,10 Various methods have already been developed to detect and quantify * To whom correspondence should be addressed. E-mail: [email protected]. (1) Background Document for Development of WHO Guidelines for DrinkingWater Quality; World Health Organization: Geneva, Switzerland, 2007. (2) Koenig, R. Science 2000, 287, 1737–1738. (3) Solomonson, L. P. In Cyanide in Biology; Vennesland, B., Conn, E. E., Knowles, C. J., Westley, J., Wissing, F., Eds.; Academic Press: London, 1981; pp 11-28. (4) Cooke, R. D.; Coursey, D. G. In Cyanide in Biology; Vennesland, B., Conn, E. E., Knowles, C. J., Westley, J., Wissing, F., Eds.; Academic Press: London, 1981; pp 93-114. (5) Padmaja, G. Crit. Rev. Food Sci. 1995, 35, 299–339. (6) Nhassico, D.; Muquingue, H.; Cliff, J.; Cumbana, A.; Bradbury, J. H. J. Sci. Food Agric. 2008, 88, 2043–2049. (7) Zelder, F. H.; Ma¨nnel-Croise, C. Chimia 2009, 63, 58–62. (8) Oluwole, O. S. A.; Onabolu, A. O.; Mtunda, K.; Mlingi, N. J. Food Compos. Anal. 2007, 20, 559–567. (9) Montagnac, J. A.; Davis, C. R.; Tanumihardjo, S. A. Compr. Rev. Food Sci. F. 2009, 8, 17–27. (10) Cardoso, A. P.; Mirione, E.; Ernesto, M.; Massaza, F.; Cliff, J.; Haque, M. R.; Bradbury, J. H. J. Food Compos. Anal. 2005, 18, 451–460. 10.1021/ac901977u CCC: $40.75  2009 American Chemical Society Published on Web 10/20/2009

endogenous biological cyanide. They include either the direct estimation of cyanogenic glycosides by gas chromatography11 or the detection of glucose or hydrogen cyanide with potentiometric,12,13 amperometric,14,15 fluorometric,16 or enzymatic techniques.17-19 Shortcomings include the time-consuming sample preparation and reaction time, the instability of the sensor system, as well as the need of expensive and bulky laboratory systems that are not available everywhere. Therefore, spectrophotometric methods have attracted much attention.20-24 Most widely used for the detection of biological cyanide are still organic-based sensors like picric acid25-27 or the so-called Koenig reagent, a combination of chloramine-T and isonicotinic acid/barbituric acid.28 Drawbacks of these sensors include their low solubility in water, the relative long reaction time, a complex multistep reaction mechanism, as well as the toxicity or explosiveness of the reagents.25 Besides, organic-based sensors were not yet successful in monitoring enzymatic reactions that liberate cyanide from their biological precursors.29 Water-soluble transition-metal complexes with high affinity to cyanide could be promising alternatives. Since the absorption spectra of metalloporphyrins and corrins are strongly dependent (11) Nahrstedt, L. P. In Cyanide in Biology; Vennesland, B., Conn, E. E., Knowles, C. J., Westley, J., Wissing, F., Eds.; Academic Press: London, 1981; pp 145181. (12) Yeoh, H. H. Biotechnol. Tech. 1993, 7, 761–764. (13) Yeoh, H. H.; Truong, V. D. Food Chem. 1993, 47, 295–298. (14) Tatsuma, T.; Tani, K.; Oyama, N.; Yeoh, H. H. J. Electroanal. Chem. 1996, 407, 155–159. (15) Tatsuma, T.; Tani, K.; Oyama, N.; Yeoh, H. H. Anal. Chem. 1996, 68, 2946– 2950. (16) Chung, S.-Y.; Nam, S.-W.; Lim, J.; Park, S.; Yoon, J. Chem. Commun. 2009, 2866–2868. (17) Albery, W. J.; Cass, A. E. G.; Shu, Z. X. Biosens. Bioelectron. 1990, 5, 367– 378. (18) Amine, A.; Alafandy, M.; Kauffmann, J. M.; Pekli, M. N. Anal. Chem. 1995, 67, 2822–2827. (19) Smit, M. H.; Rechnitz, G. A. Anal. Chem. 1993, 65, 380–385. (20) Martinez-Manez, R.; Sancenon, F. Chem. Rev. 2003, 103, 4419–4476. (21) Snowden, T. S.; Anslyn, E. V. Curr. Opin. Chem. Biol. 1999, 3, 740–746. (22) Schmidtchen, F. P.; Berger, M. Chem. Rev. 1997, 97, 1609–1646. (23) deSilva, A. P.; Gunaratne, H. Q. N.; Gunnlaugsson, T.; Huxley, A. J. M.; McCoy, C. P.; Rademacher, J. T.; Rice, T. E. Chem. Rev. 1997, 97, 1515– 1566. (24) Beer, P. D.; Gale, P. A. Angew. Chem., Int. Ed. 2001, 40, 486–516. (25) Bradbury, M. G.; Egan, S. V.; Bradbury, J. H. J. Sci. Food Agric. 1999, 79, 593–601. (26) Bradbury, J. H. Food Chem. 2009, 113, 1329–1333. (27) Brandl, H.; Lehmann, S.; Faramarzi, M. A.; Martinelli, D. Hydrometallurgy 2008, 94, 14–17. (28) Essers, S.; Bosveld, M.; Vandergrift, R. M.; Voragen, A. G. J. J. Sci. Food Agric. 1993, 63, 287–296. (29) Guilbault, G. G.; Kramer, D. N. Anal. Chem. 1966, 38, 834–836.

Analytical Chemistry, Vol. 81, No. 22, November 15, 2009

9493

on the nature of additional axial ligands,30-33 they can be utilized as analytical sensors.34-39 Recently, we communicated a corrinbased receptor as one of the few examples that detect levels of parts per million of cyanide.34,40-46 Most of these highly sensitive systems work only in the presence of organic solvents, not appropriate to study the enzymatic release of cyanide under biological conditions. In this paper, we present considerable improvements in the detection of biological cyanide. We report about the selective and rapid colorimetric detection of endogenous biological cyanide with Co(III) metal complexes. The methodology works directly in the crude biological matrix and does not depend on analytical instrumentation, time-consuming sample pretreatment, or the use of special reaction conditions. It was possible to study the enzymatic liberation of cyanide from its biological precursor directly on a freshly sliced biological surface with diffuse reflectance UV-visible (DRUV-vis) spectroscopy. EXPERIMENTAL SECTION Materials. General Information. Waxed cassava roots (Manihot esculenta Crantz) imported from Costa Rica were purchased at local supermarkets in Zurich, Switzerland. They were either freshly used or tightly wrapped in plastic wrap and stored at 4 °C until use. For safety reasons, the preparations of the cassava extracts and the acid hydrolysis were performed under a ventilated fume hood and an HCN detector from Dra¨ger was used. Potassium cyanide, dicyanocobyrinic acid heptamethyl ester (3-CN) and Ches were obtained from Fluka (Buchs, CH). Aquocyanocobyrinic acid 2 and aquocyanocobyrinic acid heptamethyl ester 3 were synthesized from their corresponding dicyano forms 2-CN and 3-CN as pairs of diastereomers through nonselective displacement of either the upper β- or the R-cyanide as described elsewhere.33,34 KCN stock solutions (10-3 M) were freshly prepared before use. The desired pH value of a stock solution of the buffer Ches (0.1 M; pH 9.5) was adjusted by the addition of either a solution of NaOH (2 M) or HCl (1 M). All measurements were performed at a final buffer concentration of 20 mM. (30) Smith, K. M. In Porphyrins Metalloporphyrins; Smith, K. M., Ed.; Elsevier: Amsterdam, 1975; pp 3-28. (31) Pratt, J. M., Ed. Inorganic Chemistry of Vitamin B12; Academic Press: New York, 1972. (32) Smith, R. G. J. Am. Chem. Soc. 1929, 51, 1171–1174. (33) Werthemann, L. PhD Thesis, ETH, Zurich, Switzerland, 1968. (34) Ma¨nnel-Croise, C.; Zelder, F. Inorg. Chem. 2009, 48, 1272–1274. (35) Poland, K.; Topoglidis, E.; Durrant, J. R.; Palomares, E. Inorg. Chem. Commun. 2006, 9, 1239–1242. (36) Freeman, M. K.; Bachas, L. G. Anal. Chim. Acta 1990, 241, 119–125. (37) Daunert, S.; Bachas, L. G. Anal. Chem. 1989, 61, 499–503. (38) Hassan, S. S. M.; Hamza, M. S. A.; Kelany, A. E. Talanta 2007, 71, 1088– 1095. (39) Zelder, F. H. Inorg. Chem. 2008, 47, 1264–1266. (40) Kim, Y.; Zhao, H.; Gabbai, F. P. Angew. Chem., Int. Ed. 2009, 48, 4957– 4960. (41) Shang, L.; Dong, S. J. Anal. Chem. 2009, 81, 1465–1470. (42) Touceda-Varela, A.; Stevenson, E. I.; Galve-Gasion, J. A.; Dryden, D. T. F.; Mareque-Rivas, J. C. Chem. Commun. 2008, 1998–2000. (43) Cho, D.-G.; Kim, J. H.; Sessler, J. L. J. Am. Chem. Soc. 2008, 130, 12163– 12167. (44) Lou, X.; Zhang, L.; Qin, J.; Li, Z. Chem. Commun. 2008, 5848–5850. (45) Niu, H. T.; Su, D. D.; Jiang, X. L.; Yang, W. Z.; Yin, Z. M.; He, J. Q.; Cheng, J. P. Org. Biomol. Chem. 2008, 6, 3038–3040. (46) Badugu, R.; Lakowicz, J. R.; Geddes, C. D. J. Am. Chem. Soc. 2005, 127, 3635–3641.

9494

Analytical Chemistry, Vol. 81, No. 22, November 15, 2009

Preparation of the Cassava Root Extract. Sectors of the fresh peeled roots were ground with a zest grater and subsequently homogenized with a pestle and mortar. Before further processing, the cassava extract was stored in a sealed tube for 60 min at room temperature in order to hydrolytically cleave the linamarin with endogenous enzymes. Previously, it was verified that no longer reaction times are necessary. Usually 2 g of finely ground material was diluted with 5 mL of water and centrifuged for 10 min at 6000 U. Aliquots of 10-100 µL of the supernatant were used for further analysis. Acid Hydrolysis Methodology. The acid hydrolysis methodology was performed as described by Haque and Bradbury47,48 as an alternative for the enzymatic release of total cyanide from cyanogenic glycosides. Therefore, 2 g of pestled biological material was diluted to 5 mL with phosphoric acid (0.1 M). The mixture was centrifuged for 10 min at 6000 U and 2 mL of the supernatant solution was added to 2 mL of sulfuric acid (4 M). It was heated at 100 °C for 90 min in a tightly closed heavy-walled Schlenk flask with a Teflon plug valve. After cooling the reaction mixture to 4 °C, the pH was adjusted to 7 with sodium hydroxide (3.6 M). Aliquots between 10 and 100 µL were used for further analysis, as described in Spectroscopic Measurements. Spectroscopic Measurements. UV-vis spectra were measured at T ) 21 ± 1 °C with a Cary 50 spectrophotometer using quartz cells with a path length of 1 cm. If not indicated otherwise, all analysis was performed with chemosensor 3 (40 µM) in water at pH 9.5 ([Ches] ) 20 mM). The UV-vis spectra were recorded in general 10 min after the addition of the aliquot of the biological extract. Previously, it was verified that the absorbance was constant over at least 5 h. The pH values were determined by using a Metrohm 827 pH meter. DRUV-vis spectra were recorded on a Varian Cary 500 spectrophotometer equipped with Internal DRA 110 mm. Stopped-Flow Kinetic Measurements. Kinetic measurements were carried out on a SX20 (Applied Photophysics) coupled to an online data acquisition system. All kinetic measurements were carried out under a controlled temperature of ± 0.1 °C and pseudofirst order conditions with an excess of CN- of at least 25-fold. Kinetic traces were analyzed at λmax ) 578 nm under consideration of the molar ratio of 2a and 2b of 1:1.2 as well as the extinction coefficients ε578 ) 570 and 9200 L mol-1 cm-1 for 2 and 2-CN, respectively.49 Values of ∆Hq and ∆Sq were calculated from the slopes and intercepts of plots of ln(k/T) versus 1/T. Preparation of Cassava Slices and DRUV-Vis Measurements. Slices of cassava (Ø ) 15.0 ± 0.2 mm; thickness 2.0 ± 0.2 mm; m ) 0.40 ± 0.04 g)50 were freshly prepared. Special care was taken not to contact the surface area afterward. A DRUV-vis spectrum was recorded as a reference 1 min after slicing. Afterward, 20 µL of chemosensor 2 (2.8 mM) was distributed uniformly on the surface. The kinetic measurements were started with a delay of 3 min from the time of slicing. (47) Bradbury, J. H.; Egan, S. V.; Lynch, M. J. J. Sci. Food Agric. 1991, 55, 277–290. (48) Haque, M. R.; Bradbury, J. H. Food Chem. 2002, 77, 107–114. (49) The ratio of the two diastereomers was determined from the integrals of the corresponding two singlets of the H(10) protons by 1H NMR spectroscopy. (50) The dimensions represent average values from three different slices.

Scheme 1. Schematic Representation of the Enzymatic Liberation of Cyanide from Cell Bound Linamarin 1: (A) Linamarase-Catalyzed Hydrolysis of Linamarin to Glucose and Acetone Cyanohydrin after Cell Rupture and (B) Formation of Hydrogen Cyanide and Acetone from the Corresponding Cyanohydrin

Scheme 2. Structural Formula and Schematic Representation of β-Aquo-r-cyanocobyrinic Acid 2 and β-Aquo-r-cyanocobyrinic Acid Heptamethyl Ester 3a

a

Only one diastereomer is shown.

Scheme 3. Proposed Mechanism and Observed Color Change of the Reaction of the β-Aquo, r-Cyano (left top) and β-Cyano, r-Aquo (left bottom) Diastereomers 2b and 2a with Cyanide to 2-CN

RESULTS AND DISCUSSION Kinetic Investigations. We investigated with stopped-flow techniques and UV-vis spectroscopy the associated kinetics of the fast substitution of Co(III)-bound water of chemosensor 2 with cyanide (Scheme 3) in order to determine the rate of the reaction as well as to get an insight into the underlying reaction mechanism.51-55 The binding constant K of 2 with cyanide is 5.5 (± 0.3) × 105 M-1 at pH 9.5 and 21 °C.34 Chemosensor 2 is available as a pair of diastereomers of the R-aquo, β-cyano (2a) and β-aquo, R-cyano (2b) forms. For the kinetic investigations, a sample with a constant molar ratio of 1:1.2 between 15 and 45 °C was used.49 Upon reaction with cyanide, both diastere(51) Prinsloo, F. F.; Meier, M.; van Eldik, R. Inorg. Chem. 1994, 33, 900–904. (52) Hamza, M. S. A.; Zou, X.; Brown, K. L.; van Edik, R. Inorg. Chem. 2001, 40, 5440–5447. (53) Illner, P.; Kern, S.; Begel, S.; van Eldik, R. Chem. Commun. 2007, 4803– 4805. (54) Wang, R.; MacGillivray, B. C.; Macartney, D. H. Dalton Trans. 2009, 3584– 3589. (55) Marques, H. M. J. Chem. Soc. Dalton 1991, 1437–1442.

Figure 1. Plot of kobsI vs concentration of cyanide for the reaction depicted in Scheme 3 as a function of temperature. Experimental conditions: pH ) 9.5 ([Ches] ) 20 mM), [2] ) 40 µM.

omers 2a and 2b convert to the same product, 2-CN, accompanied by a large bathochromic shift (∆λmax ) 51 nm) (Scheme 3). All reactions were performed in water at pH 9.5 ([Ches] ) 20 mM) under the same conditions as applied for the cyanide-sensing experiments. The kinetic measurements were run under pseudofirst-order conditions with cyanide in excess. A typical absorption vs time trace shows that the reaction is complete within 3 s.56 All traces could be fitted with a biexponential function under pseudo-first-order conditions describing the parallel reaction of two starting materials into a single product as represented by the reaction of cyanide with the diastereomeric mixture of 2a and 2b (Scheme 3).57 Values of kobsI and kobsII were measured at different concentrations of cyanide and different temperatures. The slopes of the lines in Figure 1 represent the second-order rate constants at different temperatures for the substitution of water with cyanide for one isomer of 2, as illustrated in Scheme 3. We assume that the second-order rate constant kII of 15600 ± 400 M-1 s-1 at 25 °C represents the substitution at the less crowded β-side of chemosensor 2, whereas kI of 4220 ± 50 M-1 s-1 attributes to ka of the reaction at the sterically more hindered R-side (Table 1). However, since the different diastereomers could not be separated, an unambiguous assignment is not possible. The activation parameters ∆Hq and ∆Sq were determined from the temperature dependence, for which the Eyring plot is depicted in Figure 2. The negative activation entropy of the cyanide coordination to chemosensor 2a suggests an associatively activated reaction mechanism (Table 1). (56) See Supporting Information, Figure S3. (57) Espenson, J. H. Chemical Kinetics and Reaction Mechanisms; McGraw-Hill: New York, 1995.

Analytical Chemistry, Vol. 81, No. 22, November 15, 2009

9495

Figure 2. Temperature dependence of kI and kII for the reactions of 2a and 2b with cyanide, as depicted in Scheme 3. Table 1. Rate Constants and Activation Parameters for the Coordination of CN- to 2a (I) and 2b (II)a rate constant/activation parameter kb (M-1 s-1) ∆Hq (kJ/mol) ∆Sq (J/mol K) b

I

II

4220 ± 50 45 ± 1 -25 ± 1

15600 ± 400 50 ± 2 4±5

Figure 3. 1H NMR in CDCl3 of: commercially available 3-CN (A) and 3-CN extracted from cassava (B). * ) Impurity.

a An unambiguous assignment of the diastereomers is not possible. At 25 °C.

Cyanide Liberation and Its Interaction with the Chemosensor. Cassava was chosen as a typical source for endogenous biological cyanide. The extracts were prepared and analyzed with chemosensor 2 as described in the Experimental Section. The absorption spectrum of 2-CN with absorption maxima at 367, 537, and 578 nm is identical to the spectra generated from endogenous biological cyanide and 2.58 The selectivity of the reaction was further studied with chemosensor 3 and 1H NMR-spectroscopy. This chemosensor contains the same Co(III) binding site and has the same spectrochemical properties as 2 but allows a better separation from the crude biological sample due to its increased lipophilicity.59 The corresponding 1H NMR spectrum after isolation (Figure 3B) is in good agreement with a spectrum of commercially available 3-CN (Figure 3A). The proton at H(10) is highly sensitive to the nature of additional axial ligands.33 The corresponding singlet at δ ) 5.59 ppm underscores the selective axial coordination of cyanide to the Co(III) center of 3 without any side reaction (Figure 3). Study of Enzymatic Liberation of Cyanide at a Biological Interface. The cobalt-based chemosensor 2 (Figure 4a) permits the optical detection of liberated cyanide in a diluted crude aqueous cassava solution (Figure 4b) as well as directly on a freshly ground cassava extract due to the formation of the violetcolored 2-CN complex (Figure 4c). After only four washings of the extract, no cyanide can be observed anymore with the “nakedeye” (Figure 4d).60 Besides, chemosensor 2 (1 mM) allowed also the detection of enzymatically liberated cyanide directly on the surface of a (58) See Supporting Information, Figure S1, right. (59) Ernst, L.; Maag, H. Liebigs Ann. 1996, 323–326. (60) See Supporting Information, Figure S1, left.

9496

Analytical Chemistry, Vol. 81, No. 22, November 15, 2009

Figure 4. Chemosensor 2 (1 mM) (a) applied to a diluted crude aqueous cassava solution (b), a ground cassava sample before (c) and after four washings (d), and a freshly sliced cassava (e).

freshly cut slice of cassava (Figure 4e). The sample preparation and the subsequent analysis were performed as described in the Experimental Section. The amount of cyanide liberated during 2 h from the two surfaces of a standardized slice into 2 mL of water is 1.8 ± 0.2 mg/kg.61 The UV-vis spectra of 2-CN in water with absorption maxima for the R- and β-band at λmax ) 537 and 578 nm corresponds to the DRUV-vis spectra of 2-CN on the surface of the biological sample.62 The course of the enzymatic liberation of cyanide from linamarin was followed by DRUV-vis spectroscopy directly on the surface of a freshly cut slice of cassava (Figure 5). An excess of chemosensor 2 (56 nmol) compared to the concentration of total cyanide (14 nmol) liberated from one surface of a standardized slice was used in these experiments, and the spectra were recorded every 2 min. Figure (61) The data are average values from three different slices. (62) See Supporting Information, Figure S2, right.

Figure 5. DRUV-vis spectra recorded every 2 min for the reaction of 2 (56 nmol) with enzymatically liberated cyanide on the surface of a standardized slice of cassava.

Figure 7. Relative decrease of cyanide during stirring of 1 g of finely ground cassava in 8 mL of water at 100 °C (bottom). Corresponding solutions of 2 (40 µM) in water (Vtotal ) 1 mL; [Ches] ) 20 mM; pH ) 9.5 ( 0.1) after the addition of centrifuged aliquots (30 µL) from the reaction mixture. From left to right: after 0, 3, 10, 30, and 150 min. Chemosensor 2 (40 µM) is indicated as a reference (top). Table 2. Contents of Total Cyanide of Different Biological Samples cyanide (mg/kg) entrya

sample

enzymatic hydrolysis

1 2 3

cassava 1 cassava 1 cassava 2

122 ± 3 187 ± 5 265 ± 5

a

Figure 6. Typical kinetic trace at λmax ) 578 nm for the reaction depicted in Figure 5.

5 shows the shift of the reflectance values expressed as log 1/reflectance from λmax ) 353, 497, and 527 nm to λmax ) 365, 534, and 578 nm, in good agreement with the corresponding absorption spectra of 2 and 2-CN in pure homogeneous aqueous conditions.63 Since the coordination of cyanide to chemosensor 2 is fast as analyzed with stopped-flow kinetic measurements, the formation of 2-CN reflects the degradation of linamarin to cyanide catalyzed with endogenous enzymes in real time (Scheme 1). An exponential curve described by a first-order kinetics with a half-life of 12.5 (± 0.2) min fits remarkably accurately with the observed kinetic trace (Figure 6), but the diffusion of 2 into the biological matrix excludes a detailed quantitative analysis. Cyanide Quantification. The content of total cyanide of the biological samples was estimated spectrophotometrically at λmax ) 367 and 578 nm by comparison with a calibration curve generated from the titration of 2 (40 µM) with cyanide (2-45 µM) in water ([Ches] ) 20 mM, pH 9.5).64 Since the enzymatically liberated cyanohydrins convert rapidly to the corresponding acetone and hydrogen cyanide under basic conditions, all spectroscopic investigations were carried out at pH 9.5 ([Ches] ) 20 mM).4 Table 2 summarizes the contents of total cyanide for two different cassava roots determined by two independent methods. First, the contents of total cyanide of the finely ground biological samples were determined directly with chemosensor 2 (Table 2, column 3). The results are in good agreement with the determination of cyanide following the acid (63) See Supporting Information, Figure S2. (64) See Supporting Information.

acid hydrolysis 125 ±9 184 ± 9 259 ± 10

The data are averaged values from at least two measurements.

hydrolysis47 (Table 2, column 4), indicating that endogenous enzymes sufficiently hydrolyze the cyanogenic glycosides.48 The contents of total cyanide varied between different roots from 122 ± 3 to 265 ± 5 mg/kg (Table 2, entries 1, 3) and also within one cultivar (∆conc ) 65 ± 8 mg/kg; Table 2, entries 1, 2).4 Cyanide Detection during Food Processing. Two test series were conducted to demonstrate the general applicability of chemosensor 2 as a fast and simple detection method for the determination of total cyanide during food processing. The removal of hydrogen cyanide through boiling was compared to the washing of a cassava extract, two different methods that are usually applied in combination in cassava processing.5 Figure 7 represents the decrease of cyanide during stirring of 1 g of finely ground cassava in 8 mL of water at 100 °C. Although, 45% of the cyanide is released in the first 10 min, further boiling has only a minor influence, suggesting that the enzyme linamarin denaturates at elevated temperatures. This can also be observed qualitatively by a change from the violet-colored 2-CN at the beginning of the experiment to the rose-violet-colored mixture of 2-CN and 2 at advanced reaction times. On the other hand, washing removes cyanide more efficiently. After only four washings, no cyanide could be detected anymore in the sample with chemosensor 2.60 CONCLUSION Aquocyanocorrinoids allow the selective and rapid detection of endogenous biological cyanide in colorless plant samples and extracts. The methodology works without time-consuming sample pretreatment and the use of special reaction conditions. Therefore, it was possible to study the course of the enzymatic liberation Analytical Chemistry, Vol. 81, No. 22, November 15, 2009

9497

directly on a freshly sliced biological surface with diffuse reflectance UV-visible (DRUV-vis) spectroscopy. This form of analysis gives first-hand insight into the self-defense mechanism of more than 2000 plants that liberate cyanide in order to protect themselves from animals and insects. The investigations were accompanied by stopped-flow kinetic measurements under homogeneous conditions. Besides, cyanide detection can be readily understood in the form of a simple yes-no selectivity as a result of a color change from orange to violet. Possible applications during food manufacturing are demonstrated. Compared to organic-based sensors like the widely established picric acid test, corrin-based chemosensors could provide an important improvement in food safety control. It is expected that this methodology will find numerous further applications in the near future.

9498

Analytical Chemistry, Vol. 81, No. 22, November 15, 2009

ACKNOWLEDGMENT C.M.-C. is grateful for the support of the Forschungskredit of the University of Zu¨rich. The authors thank A. Szentkuti for assistance and R. Alberto for helpful discussions.

SUPPORTING INFORMATION AVAILABLE Additional information as noted in the text. This material is available free of charge via the Internet at http://pubs.acs.org.

Received for review September 2, 2009. Accepted October 9, 2009. AC901977U