Cyanide detection in gastric juice with corrin-based chemosensors

Cyanide (CN ) is highly toxic to humans and most forms of life because of its strong binding to biologically active transition metal ions. In particular, it effectively blocks the vacant coordination sites of ferri-(Fe) containing heme complexes and thereby inactivates cellular respiration. Cyanide intoxications may occur from intentional and accidental cyanide uptake and are most frequently encountered in industrialized countries aer smoke inhalations from res. Indeed, more than 80% of re victims die from smoke gas intoxications and not from burns. One of the most toxic combustion products of nitrogen containing organic products is hydrogen cyanide (HCN; pKa 1⁄4 9.04 (ref. 8)), the conjugated acid of cyanide. Testing for cyanide intoxication is therefore required in clinical diagnostics and forensic investigations. For this purpose, different types of body uids such as blood, saliva and gastric acid are usually analysed. Most methods require sample pre-treatment, sophisticated technical equipment (e.g. GC-MS, ESI-MS-MS) as well as specialised users and hence, the analysis of cyanide is not yet routine in hospital and forensic settings. Recent developments indicate that low-cost and straightforward test-kits with colorimetric and uorometric indicators may overcome these limitations. In this context, vitamin B12 based chemosensors are currently representing one of the most promising systems for detecting cyanide in biological samples in terms of handling, selectivity, sensitivity and response time. Based on earlier pioneering investigations of our group, we describe herein the detection of cyanide in articial gastric juice with corrin-based chemosensors and solid phase extraction (SPE). The applicability of the method is demonstrated in a proof-of-concept forensic investigation.


Introduction
Cyanide (CN À ) is highly toxic to humans and most forms of life because of its strong binding to biologically active transition metal ions. 1,2 In particular, it effectively blocks the vacant coordination sites of ferri-(Fe III ) containing heme complexes and thereby inactivates cellular respiration. [3][4][5] Cyanide intoxications may occur from intentional and accidental cyanide uptake and are most frequently encountered in industrialized countries aer smoke inhalations from res. 6 Indeed, more than 80% of re victims die from smoke gas intoxications and not from burns. 7 One of the most toxic combustion products of nitrogen containing organic products is hydrogen cyanide (HCN; pK a ¼ 9.04 (ref. 8)), 9 the conjugated acid of cyanide. Testing for cyanide intoxication is therefore required in clinical diagnostics and forensic investigations. 1,2,10,11 For this purpose, different types of body uids such as blood, 12-17 saliva 18,19 and gastric acid 20,21 are usually analysed. Most methods require sample pre-treatment, sophisticated technical equipment (e.g. GC-MS, 22 ESI-MS-MS 23 ) as well as specialised users and hence, the analysis of cyanide is not yet routine in hospital and forensic settings. 2 Recent developments indicate that low-cost and straightforward test-kits with colorimetric and uorometric indicators may overcome these limitations. 12,17,[24][25][26][27][28][29] In this context, vitamin B 12 based chemosensors are currently representing one of the most promising systems for detecting cyanide in biological samples in terms of handling, selectivity, sensitivity and response time. 4,26,[28][29][30][31][32][33][34] Based on earlier pioneering investigations of our group, [34][35][36][37] we describe herein the detection of cyanide in articial gastric juice with corrin-based chemosensors and solid phase extraction (SPE). 24,38 The applicability of the method is demonstrated in a proof-of-concept forensic investigation.
UV-vis spectra were measured at T ¼ 21 AE 1 C with a Cary 50 spectrometer (Varian, Switzerland) using quartz cells with a path length of 1 cm.

Preparation of solutions
The desired pH value of the stock solution of the buffer Ches (0.25 mol L À1 ; pH 9.6) was adjusted by the addition of 2 mol L À1 NaOH and 1 mol L À1 HCl solutions. The pH values were measured with a Metrohm 827 pH lab (Metrohm, Switzerland).
The cyanide stock solution (10 À3 mol L À1 ) was freshly prepared every day.
Cyanide was either detected directly in solution (A) or aer solid-phase extraction (B).
A: The solution was transferred to a quartz cuvette and analyzed with UV-vis spectroscopy.
B: Aer one minute, the sample was passed through a preconditioned C18ec column and subsequently washed with water (3 mL) to remove adhering gastric juice from the C18ec material. The chemosensor was subsequently eluted from the C18ec column with MeOH (0.4 mL). The eluate volume was adjusted to 0.5 mL with MeOH, transferred to a quartz cuvette and analysed with UV-vis. The cyanide content was determined with the help of a calibration curve.

Calibration curves
Homogenous conditions. UV-vis spectra of AGJ samples spiked with different amounts of cyanide ([CN À ] ¼ 0-160 mmol L À1 ) were analysed. The calibration curve was obtained by plotting the absorptions at 583 nm against CN À concentrations. Each experiment was repeated three times to assign mean values and standard deviations. A linear equation (y ¼ a + bx) was obtained by tting the linear range with OriginPro2015.
Solid-phase extraction. UV-vis spectra of AGJ samples spiked with different amounts of cyanide ([CN À ] ¼ 0-160 mmol L À1 ) were analysed aer eluting the reagent from the C18ec cartridge with MeOH (V ¼ 0.4 mL). The eluate volume was adjusted to 0.5 mL with MeOH and the UV-vis spectra were recorded. The calibration curve was obtained by plotting the absorptions at 583 nm against CN À concentrations. Each experiment was performed three times in order to determine the mean value and standard deviations. A linear equation (y ¼ a + bx) was obtained by tting the linear range with OriginPro2015.

Articial gastric juice cyanide detection at different incubation times
The inuence of incubation time (dened as time between spiking and analysis) was determined for AGJ samples stored at room temperature (T ¼ 23 AE 1 C). Samples spiked with cyanide (V sample ¼ 0.5 mL, [CN À ] AGJ ¼ 40 mmol L À1 ) were analysed aer different time intervals (t ¼ 20-1470 min) and analysed as described under Section 2.4. Each experiment was repeated three times to assign mean values and standard deviations. In the same way, experiments were performed for AGJ samples

Gastric juice cyanide detection in authentic human samples
Cyanide was monitored in diluted human gastric juice samples (HGJ) with (A) and without (B) solid-phase extraction using corrin-based chemosensors. In addition, the samples were analysed by a microdiffusion assay with barbituric acid/pyridine as reagent (C). 41 Each experiment was either performed three times (A, B) or twice (C) in order to determine the mean value and standard deviations.
A: Ches buffer (472 mL, 0.25 mol L À1 ) and ACCbs (28 mL, 1.5 mmol L À1 , 42 nmol or 42 mL, 1 mmol L À1 , 42 nmol) were added to a HGJ sample (V sample ¼ 0.5 mL). Aer one minute, the sample was passed through a Chromabond C18ec column. The column was washed with water (3 mL). Immobilized corrinoids were eluted with MeOH (0.4 mL). The eluate volume was adjusted to 0.5 mL, transferred to a quartz cuvette and analyzed with UV-vis. The concentration of cyanide in HGJ was determined with the help of the calibration curve shown in Fig. 4. B: Ches buffer (200 mL, 0.25 mol L À1 ) and ACCbs (42 mL, 1 mmol L À1 , 42 nmol) were diluted with water (758 mL, V sample ¼ 1 mL). Aliquots (10-100 mL) of diluted HGJ (dilution factor: 10) were added. Aer 5 minutes, the concentration of cyanide in HGJ was determined with the help of the calibration curve shown in Fig. 3 aer volume correction and under considering of the dilution factor. C: Sulfuric acid (10%, 0.5 mL) and diluted gastric juice (1 mL, 1 : 50 and 1 : 100 dilution with distilled water, respectively) were added to the outer compartment of two Conway microdiffusion cells. To the inner compartment sodium hydroxide (1 M, 0.5 mL) was added. The cells were immediately closed, gently agitated and incubated for 2 hours. An aliquot (0.1 mL) of the inner compartment was transferred into a tube containing a hydrogen phosphate solution (0.87 mol L À1 , 1 mL). A chloramine-T solution (25%, 0.5 mL) was added and incubated for 3 min at room temperature. Finally, a barbituric acid/pyridine reagent (1.5 mL; reagent: 3 g of barbituric acid in 15 mL of pyridine and hydrochloric acid (32%, 3 mL) were diluted with distilled water to a nal volume of 25 mL) was added. It was incubated for another 10 min and the absorbance of the solution was measured at 580 nm. Concentrations were calculated by comparison with a calibration curve using standard cyanide solutions prepared in the same manner.  The hypsochromic shi of the g-band of ACCbs from 355 nm to 359 nm and the broadening of the a-, and b-bands (497 and 527 nm) to one single absorption (525 nm) indicates interactions between the chemosensor and AGJ. Most likely, an amino acid side chain of the pepsin, the main component of AGJ, is coordinated to the vacant coordination site of the chemosensor ( Fig. 2A). 42,43 Titrating cyanide ([CN À ] ¼ 0-160 mmol L À1 ) to these solutions led to the stepwise formation of the dicyano-form of the indicator (Fig. 2 and 3). This behaviour is indicated by shis of the g-band from 359 nm to 368 nm and of the a/b-band from 526 nm to 583 nm. These characteristic absorptions are identical with those of dicyanocobester in water. 36,44 Saturation of the chemosensor (42 mmol L À1 ) with cyanide is reached in AGJ containing solutions at $110 mmol L À1 . This concentration is approximately twice as high as for the same experiment in pure water (pH 7.5). 36 This difference can be rationalized by the competitive binding of pepsin to the chemosensor ( Fig. 2A) that makes coordination of cyanide more challenging (Fig. 2C) compared to the reaction of cyanide with the 'free' chemosensor (Fig. 2B).

Results and discussion
Aer having established the detection of cyanide in AGJ under homogenous conditions, we evaluated the possibility to combine the corrin-based chemosensors with solid phase extraction. 24,26,27,45 For this purpose, the reagent was added to a cyanide spiked AGJ sample and extracted aerward on the top of a commercially available reverse phase extraction column as described in the experimental section. The chemosensor is thereby trapped on the top of the C18ec column and is visible as a coloured ring. Depending on the concentration of cyanide, its colour varies between orange (aquacyano-form) and violet (dicyano-form). 27,36 For quantitative determinations, the reagent was subsequently eluted from the column with MeOH and analysed with UV-vis spectroscopy. The spectra unambiguously indicate a shi of the chemosensor (l g-band ¼ 359 nm; l a/b-band ¼ 526 nm) to its dicyano-form (l g-band ¼ 368 nm; l a-band ¼ 583 nm) with increasing concentrations of cyanide (Fig. 4 right). Simultaneously, strongly pronounced increases in absorbance are observed. For example, the intensity of the g-band increases 2.4 times during titrations (Fig. 4 right), whereas it changes solely by $20% in the same experiment under homogenous conditions (Fig. 3 le). This behaviour has been observed earlier and is explained by the more efficient extraction of the neutral dicyano-compound compared to the more polar positively charged aquacyano-form during SPE. 24 The corresponding calibration curve for cyanide shows a linear range up to 120 mmol L À1 (Fig. 4).
Having demonstrated that cyanide detection is possible in AGJ with corrin-based chemosensors, we studied potential inuences of incubation times and modes of sample storage.
Cyanide spiked samples of AGJ were stored at room temperature for up to 24 hours and analysed at different time intervals. No signicant deviations were observed indicating that 'free' cyanide is sufficiently stable in AGJ under these conditions (Fig. 5). Storage at À21 C was also tested (t ¼ 24-1320 h). In this series of experiments, the detectable amount of cyanide dropped by 29 AE 5% aer one week of storage (data not shown). Surprisingly, longer storage times did not show any additional negative effect and cyanide can be still detected in AGJ with corrin-based chemosensors and SPE aer 55 days of storage at À21 C.
Aer having established the corrin-based SPE method for cyanide detection in AGJ, the system was tested in a proof-ofconcept forensic study. An anonymous sample from a suicide case, in which the victim had swallowed a solution of potassium cyanide, was evaluated. In contrast to the colorless AGJ sample, the human gastric juice sample (HGJ) was slightly reddish indicating the presence of additional, unknown components. Samples of HGJ were therefore rst tested with SPE to avoid any undesired colour interferences from the HGJ sample. 27 Indeed, a clear color change of the chemosensor from orange to violet was observed while testing the HGJ samples (Fig. 6 le). This behaviour suggests the presence of free cyanide in the HGJ sample as supported by UV-vis spectroscopy aer eluting the reagent from the column (Fig. 6 right). The absorbance spectrum of the eluted compound indicated unambiguously coordination of cyanide to the chemosensor generating its corresponding dicyano-form. 36 Quantication of cyanide in HGJ was then performed with the help of the previously established calibration curve (Fig. 4).
The analysis of a 267 times diluted HGJ sample indicated a cyanide concentration of 11 900 AE 900 mmol L À1 . This result is in very good agreement with an independent microdiffusion assay using barbituric acid/pyridine as reagent. 41 This assay showed a cyanide concentration of $10 800 AE 100 mmol L À1 in HGJ of the suicide victim. Apart from this quantitative accordance, it is important to note that cyanide was still detected by 'naked-eye' in a $1000 times diluted sample with corrin-based chemosensors and SPE.
In another set of experiments, we tested whether the extraction step of the chemosensor from the HGJ sample could be omitted and tried to directly determine the concentration of cyanide in HGJ with the corrin-based reagent. However, although the formation of the corresponding dicyano-form of the chemosensor could by easily detected in the diluted HGJ sample (dilution factor: 969), quantication led to larger deviations compared to the SPE technique. For example, we calculated a cyanide concentration in HGJ of 16 100 AE 800 mmol L À1 instead of $11 000 mmol L À1 . Apparently, the calibration curve generated from cyanide spiked AGJ (Fig. 3) cannot be used for this experimental set-up. Indeed, we demonstrated in this publication that the overestimation can be avoided by separating the reagent from the HGJ sample with SPE. This behaviour suggests the existence of additional (unknown) interferents in HGJ compared to AGJ that are removed during washings (see experimental part) with the SPE technique.
To the best of our knowledge, this technical note describes the unprecedented detection of cyanide in HGJ with chromogenic chemosensors. In particular, the results obtained with corrin-based chemosensor and SPE method are in very good agreement with the microdiffusion assay using barbituric acid/ pyridine as reagent. However, the SPE method is advantageous in terms of sample preparation and speed of detection. In particular, only the conditioning of the column and sample dilution is required. Admittedly, only one authentic human sample was studied so far, but the results are encouraging for further studies in order to establish the system for routine   forensic investigations. These efforts would greatly benet from the simultaneous rapid detection of cyanide in authentic human blood samples and corresponding forensic studies are planned for the future.

Conclusions
The detection of cyanide in articial as well as authentic human gastric juice samples using corrin-based chemosensors and SPE extraction is described. The inuence of spiking times and modes of storage were evaluated. It has been demonstrated that rapid, straightforward qualitative yes/no evaluations by 'naked eye' is possible for concentrations of cyanide in HGJ as low as 10 mmol L À1 . Quantitative measurements are in very good agreement with an independent microdiffusion assay using barbituric acid/pyridine as reagent. The SPE test-kit is therefore expected to complement existing current methods, especially when rapid qualitative and quantitative analyses of cyanide in HGJ are required.