- Email: [email protected]
A new method for the determination of sulphide in gastrointestinal contents and whole blood by microdistillation and ion chromatography
Clinica Chimica Acta 293 (2000) 115–125 www.elsevier.com / locate / clinchim
A new method for the determination of sulphide in gastrointestinal contents and whole blood by microdistillation and ion chromatography Caroline J. Richardson, Elizabeth A.M. Magee, John H. Cummings* MRC Dunn Clinical Nutrition Centre, Cambridge, UK Received 1 March 1999; received in revised form 15 November 1999; accepted 17 November 1999
Abstract Hydrogen sulphide is produced in the human large intestine by the bacterial reduction of dietary inorganic sulphate and sulphite and by fermentation of sulphur amino acids. Sulphide may damage the colonic epithelium and has been implicated in the pathogenesis of ulcerative colitis. The accurate measurement of sulphide in biological samples, particularly in gut contents is difficult due to the volatile nature of the compound, and the viscosity and turbidity of the samples. Here we describe a method for the determination of sulphide in gut contents and whole blood which overcomes these problems. Initially, samples are treated with zinc acetate to trap sulphide. A microdistillation pretreatment is then used, which releases sulphide from its stable, stored state, coupled to ion chromatography with electrochemical detection. The limit of detection of the method was determined as 2.5 mmol / l, which enabled sulphide levels in gut contents and whole blood samples obtained from humans to be accurately determined. A preliminary investigation in healthy human subjects showed blood sulphide ranged from 10 to 100 mmol / l. Whole blood sulphide did not change significantly when increasing amounts of protein from meat were fed. However, faecal sulphide did show a significant increase from 164 to 754 nmol / g in four subjects fed diets which contained 60 and 420 g meat. 2000 Elsevier Science B.V. All rights reserved.
1. Introduction Hydrogen sulphide is an abundant by-product of both naturally occurring and *Corresponding author. Present address: Molecular and Cellular Pathology, Ninewells Hospital and Medical School, Dundee DDl 9SY, UK. 0009-8981 / 00 / $ – see front matter 2000 Elsevier Science B.V. All rights reserved. PII: S0009-8981( 99 )00245-4
116
C. J. Richardson et al. / Clinica Chimica Acta 293 (2000) 115 – 125
industrial processes. It has a high level of toxicity, comparable to that of hydrogen cyanide [1] and acts by the inhibition of cytochrome oxidase and thus arrests aerobic respiration, especially in nervous tissue. Several cases of fatal poisoning from H 2 S gas have been reported, and this has led to the development of a diverse range of detection methods [2]. In the human large intestine, hydrogen sulphide is produced as a product of bacterial metabolism by two known pathways. Firstly by the reduction of sulphate by sulphate-reducing bacteria (SRB) [3] and secondly by the fermentation of sulphur-containing amino acids, cysteine and methionine [4]. The presence of this toxic metabolite in the colon may be of importance in disease — in particular it has been implicated in the pathogenesis of ulcerative colitis. In vitro in both rat and human colonocytes, Roediger et al. [5,6] have shown that HS 2 at 2 mmol / l inhibits butyrate oxidation. This biochemical lesion is characteristic of the defect seen in ulcerative colitis and is due to persulphide formation which blocks butyryl CoA dehydrogenase. In perfusion studies of the rat colon, HS 2 at 0.2 mmol / l produces mucosal ulceration, goblet cell loss, apoptosis and distortion of the crypt architecture [7]. Using human colon tissue, Christl et al. [8] have shown that HS 2 at 1 mmol / l significantly increases cell proliferation rates. Furthermore, early studies of ulcerative colitis [9] showed that reliable animal models can be devised by feeding sulphated polysaccharides such as dextran sulphate and carageenan. These polysaccharides deliver sulphates directly to the colon. Bacteria are essential for this model [10]. An accurate and sensitive method for the detection of sulphide in gut contents is therefore important. The determination of volatile compounds such as sulphide in biological samples presents several obstacles. Sulphide is readily oxidised in air and in addition adheres to glass and rubber. Samples must therefore be prepared and stored in order to limit the loss of H 2 S gas. This is usually achieved by trapping the sulphide as stable zinc sulphide precipitate immediately a sample is obtained. The formation of the metal precipitate then poses problems in itself. Turbidity of the samples (in the case of gut contents) causes interference with colorimetrically determined assays such as the methylene blue, and these assays also lack sensitivity [11]. Direct analysis of samples where sulphide has been trapped as a metal complex by methods such as ion chromatography is unrealistic, since the complex formed is strong and therefore produces poor peak shapes and poor reproducibility as the complex slowly dissociates. It is therefore necessary to release the sulphide from its complex e.g. by the addition of acid and trap it in another solution. This has previously been achieved in turbid water and brain tissue samples using a gas dialysis procedure [2,12]. However, this method employs highly specialised and expensive apparatus and specially constructed gas dialysis blocks. Here, we present a microdistillation procedure in which H 2 S is first trapped as
C. J. Richardson et al. / Clinica Chimica Acta 293 (2000) 115 – 125
117
stable sulphide complexes and then distilled into a sodium hydroxide solution and analysed by ion chromatography. Preliminary results of sulphide in blood and faeces from human feeding studies are also given.
2. Method
2.1. Volunteers and diets Fresh faecal samples were collected from 15 free-living healthy volunteers. Blood was collected from six similarly free-living volunteers. Four healthy subjects, who were taking part in a controlled feeding study in the metabolic suite at the Dunn Clinical Nutrition Centre in which different amounts of meat were fed to observe the effects on faecal excretion of N-Nitroso compounds, also gave a blood sample. This was taken fasting on the final day of a 15-day diet period during which the subjects were eating either a no meat diet, or one containing 240 or 420 g meat [13]. Similarly four volunteers taking part in a controlled diet study of the effect of meat and sulphur additives in the diet on sulphur metabolism in the gut gave a faecal sample. The diets contained either 60 g meat or 420 g meat with sulphur additives.
2.2. Preparation and storage of samples All solutions used, with the exception of sodium hydroxide and eluent solutions, were deoxygenated by autoclaving at 1208C, 15 p.s.i, for 20 min, then topping up the solutions to the brim while still hot and leaving to cool tightly sealed. A 10% (w / v) faecal slurry was made by homogenising the freshly voided faecal sample in 0.1 mol / l potassium phosphate buffer pH 7.4 for 20 min in a stomacher. (Colworth Stomacher 400). The homogenate was then filtered through a 500-mm sieve to remove particulate matter, and 4 ml of filtrate was added to 1 ml of deoxygenated 120-g / l zinc acetate in a 7-ml bijou bottle. This traps the sulphide as zinc sulphide precipitate. Validation experiments using portions of a freshly passed faecal specimen showed that there was no difference in sulphide recovery if the slurry with zinc acetate was prepared as stated above or by homogenising the faecal specimen in a mixed trapping solution of anoxic phosphate buffer and zinc acetate prior to sieving. A single faecal slurry was divided into multiple aliquots and spiked and unspiked samples were stored in 7-ml bijou bottles with aluminium caps with rubber seals for up to 6 months at 48C without loss of sulphide.
118
C. J. Richardson et al. / Clinica Chimica Acta 293 (2000) 115 – 125
Whole blood was collected using 5-Monovette EDTA KE blood tubes (Sarstedt, Leicester, UK) and stored immediately at 2 208C.
2.3. Microdistillation procedure The microdistillation apparatus is shown in Fig. 1. After several prototypes had been tried, the trap units were manufactured by Soham Scientific (Ely, UK). The traps consist of two glass tubes, one for the test sample which has an injection port into which acid is injected to release hydrogen sulphide. The other tube is used to trap the hydrogen sulphide as anionic sulphide in 1 mol / l of sodium hydroxide solution. The two tubes are joined by a head unit which enables H 2 S gas evolved to be carried into the trapping solution. The design also allows the introduction of inert carrier gas directly into the sample tube, so mixing the sample throughout the procedure. The deoxygenated trapping solution was prepared by adding 40 g of sodium hydroxide pellets (BDH) to 1 l of deoxygenated water on the day of use. One ml of the trapping solution was pipetted into the trapping side of the unit and 1 ml of well-mixed sample was pipetted into the sample side of the unit. For blood, serum and other proteinaceous samples, 50 ml of a 10% (v / v) solution of antifoam SE-35 (Sigma) was added to 450 ml of 120 g / l zinc acetate and 500 ml of sample. The
Fig. 1. Microdistillation apparatus.
C. J. Richardson et al. / Clinica Chimica Acta 293 (2000) 115 – 125
119
head unit was then secured in position by sealing the joints with high vacuum silicone grease (BDH) and keck clips (Quickfit). In order to speed up the trapping process and improve recoveries, the sample side of the units were placed into a water bath at 508C, leaving the trapping side at room temperature. Increasing the temperature at which the evolution of HS 2 occurs greatly improved recoveries. However, increasing the temperature above 508C resulted in poor reproducibility and recoveries due to sample evaporation and condensation problems. Gas lines were attached to each unit and oxygen free nitrogen was bubbled through the traps at 200 cm 3 / min. After 5 min, 1 ml of concentrated HCl (BDH, Lutterworth, UK) was added to the sample via the injection port using a 21 g A2 needle (Microlance) and 1-ml syringe. The units were then left for 3 h at 508C. During assay validation recoveries of greater than 85% were achieved in 1.5 h under these conditions, but in order to increase recoveries consistently to above 85% and ensure day to day reproducibility, the time of the procedure was increased to 3 h. A volume of 300 ml of trapping solution was then transferred to a 0.5-ml polyvial (Dionex P/ N 38142) and 300 ml eluent added.
2.4. Detection of sulphide by ion chromatography Eluent was prepared by dissolving 0.6753 g oxalic acid (Sigma) in 984 ml of ultra pure water, purging with helium to degas, then adding 16 ml of 50% (w / v) solution sodium hydroxide. A Dionex DX-500 chromatography system, with a Carbopac PA-1 column 4 3 250 mm (Dionex PIN 35391) and Carbopac PA guard column 3 3 25 mm (Dionex P/ N 37141) were used with a 50-ml sample loop and 1 ml / min eluent flow-rate. A silver working electrode electrochemical cell was used (Dionex P/ N 44110), set at 0.00 V applied potential and output range 1 nA. It was found that the continuous injection of sulphide produced an oxidised precipitate Ag 2 S on the surface of the electrode. This precipitate does not effect the sensitivity of the electrode [14] though to prevent excessive build-up, the electrode surface was cleaned before each day of use, usually 80 samples, by gently rubbing with a clean pencil eraser and rinsing with ultra pure water ( . 18 MV)). The system is then allowed to equilibrate after which six injections of a 50 mmol / l sulphide standard solution are injected and reproducibility checked, as reported previously by Rocklin and Johnson [14]. Reproducibility was best if the sample vials were filled with 300 ml of sample and 300 ml of eluent, a total volume of 600 ml which was the full capacity of the vials. Calibration standards of 50, 20 and 10 mmol / l were made up in 1 mol / l sodium hydroxide on the day of analysis. Run time for a batch of 80 samples was 10 h.
120
C. J. Richardson et al. / Clinica Chimica Acta 293 (2000) 115 – 125
3. Results
3.1. Validation of ion chromatography detection The current response, due to the oxidation or reduction of solutes that occur across the electrode surface, which is measured by electrochemical detectors, is dependent upon the potential applied to the electrode (Eapp ). In this system the optimum Eapp for a 20 mmol / l ammonium sulphide standard was determined as 0 V. A calibration curve is constructed for each run. The linear range of the calibration curve was found to be 10 mmol / l to 1000 mmol / l. The limit of detection was determined as 2.5 mmol / l. Repetitive injections of a 20 mmol / l ammonium sulphide standard over 7 h gave a relative standard deviation of 3.9%. It was therefore concluded that sulphide stored in eluent was stable enough to be left in polyvials overnight in the autosampler, and this enabled us to run much larger batches of samples.
3.2. Validation of the microdistillation procedure Table 1 details the recovery of sulphide from faecal slurries which had been stored in zinc acetate and then spiked with varying concentrations of ammonium sulphide standard before the microdistillation procedure. Using this method, recoveries of 0.25, 0.5, 0.75, 1.0 and 1.5 mmol / l spikes were 98.365.3%, 101.766.9%, 94.0%, 95.563.5% and 98% respectively. Recoveries for the methylene blue assay in our laboratory ranged from 77.1 to 91.5%. Recovery experiments for the microdistillation method were also performed on the whole Table 1 Recovery of ammonium sulphide from spiked faecal slurries Sample concentration (nmol / g wet weight faeces) 6S.D.
(NH 4 ) 2 S spike concentration (nmol / g wet weight faeces)
Recovery (%)
110610 18064 15064 12606100 60065 15067 6061 72063 72062 52064 35065
1500 1000 1000 750 500 500 500 250 250 250 250
98 99 92 94 108 105 92 97 96 107 93
C. J. Richardson et al. / Clinica Chimica Acta 293 (2000) 115 – 125
121
Table 2 Recovery of ammonium sulphide from spiked whole blood samples in six subjects Blood sulphide concentration (mmol / l)6S.D
Concentration of (NH 4 ) 2 S spike (mmol / l)
Recovery (%)
2.160.03 2.660.10 3.060.10 2.560.05 5.160.59 6.0460.16
5.0 5.0 5.0 10.0 10.0 10.0
102 98 100 91 79 81
blood samples (Table 2) and were 10061.63% and 83.765.25% for the 5 and 10 mmol / l spike. Because hydrogen sulphide is such a volatile compound, it was also necessary to determine losses incurred throughout the entire procedure. This was achieved by adding a series of spikes ranging from 500 to 1000 mmol / l ammonium sulphide to freshly voided faecal samples before homogenisation. Recoveries ranged from 85.5 to 91.7%, determined using the microdistillation / IC procedure. We then compared faecal sample results obtained by this method and those obtained in parallel, on the same sample, by the methylene blue method (Table 3). A direct comparison of results obtained by the methylene blue assay and the Table 3 Comparison of sulphide values obtained from faecal slurries by the microdistillation / ion chromatography and the methylene blue assay Sulphide concentration determined by methylene blue assay (nmol / g wet weight faeces)
Sulphide concentration determined by microdistillation / IC (nmol / g wet weight faeces)
100 180 260 270 300 370 470 470 550 590 610 640 680 1700 1880
180 190 460 260 370 480 430 460 550 620 640 640 680 1200 1350
122
C. J. Richardson et al. / Clinica Chimica Acta 293 (2000) 115 – 125
microdistillation / ion chromatography method compared well between the two assays at most sample concentrations (r 2 5 0.91) although there appears to be overestimation of sulphide at concentrations over 1000 nmol / g wet weight with the methylene blue method. Fig. 2 shows the methylene blue calibration curve at concentrations above 5 mmol / l. This equates to an original sample concentration of 1000 nmol / g wet weight. Whole blood sulphide in four volunteers living on controlled diets with
Fig. 2. Methylene blue calibration curve (r 2 5 1.00).
C. J. Richardson et al. / Clinica Chimica Acta 293 (2000) 115 – 125
123
differing levels of meat intake were (mmol / l6S.E.M.) meat free 42 (15); 240 g of meat 47 (12); and 420 g of meat 25 (0.8). Faecal sulphides from a similar feeding study (nmol / g wet weight6S.E.M.) were 60 g of meat 164 (19) and 420 g of meat with sulphur additives 754 (140) P , 0.05.
4. Discussion Sulphide measurement in biological samples is difficult because of its volatility, susceptibility to oxidation, adsorption to glass and rubber and binding to organic molecules [15]. Because of this elaborate methods have been developed. The present method measures free sulphide and is simple and robust. Adsorption to glass and rubber is avoided by ensuring that samples do not come into contact with either substance prior to precipitation with zinc acetate. The addition of zinc acetate forms a very strong, stable complex preventing evolution of hydrogen sulphide. Overall recoveries from spiked fresh samples were less than 100%, but we believe that this is due to loss of H 2 S incurred during the initial homogenisation and filtration steps of sample preparation. It is therefore advisable that the time taken to prepare samples is kept to a minimum. Previously the detection of sulphide in diverse samples such as turbid and waste waters, brain tissue and serum have used a range of methods. These published methods have included the use of ion specific electrodes coupled to the Conway plate method [16], pre-column derivatisation of sulphide (based on the methylene blue assay [17] or monobromoamine derivatisation [18]) with detection by HPLC, and the use of gas dialysis with ion chromatography [2,12]. Ion specific electrodes have been reported to measure sulphide concentrations at 10 ng / ml (0.29 mmol / l), but these electrodes often give slow responses. They cannot be used directly on many biological samples because cyanide and sulphide can bind to components in the matrix. Attempts to separate cyanide and sulphide from the matrix using Conway plates have been documented, but no recovery data were published and we found poor recoveries of ammonium sulphide (0–60%) and irreproducible results from gut contents using microdiffusion plates. The use of pre-column derivatisation of sulphide by monobromoamine with HPLC separation and fluorescence detection is very sensitive (40 nmol / l) but requires very specialised equipment and it is also very labour intensive, unlike the microdistillation procedure. Faecal sulphides in a group of 15 volunteers ranged from 110 to 720 nmol / g wet weight. Feeding studies in healthy volunteers, which we have undertaken previously, have shown that faecal sulphide, measured by the methylene blue technique, can be manipulated by intake of protein from meat [19]. In these studies faecal sulphide concentrations ranged from 220 to 2800 nmol / g wet weight faeces. Other studies using the methylene blue assay have shown median
124
C. J. Richardson et al. / Clinica Chimica Acta 293 (2000) 115 – 125
(interquartile range) faecal sulphides of 250 (120–450) nmol / g in healthy controls and 680 (320–1070) nmol / g in untreated ulcerative colitis patients [20]. Using the new method we have shown in a preliminary analysis from an ongoing feeding study of healthy subjects that faecal sulphide increases significantly from 160 to 750 nmol / g when meat intake is increased, along with food containing sulphur additives as preservatives. Total protein intake in this study increased from 64 to 165 g per day. The lower detection limit of the new procedure enabled us to investigate sulphide levels in blood samples taken from volunteers on different protein diets. The data obtained showed that although sulphide could be easily detected in blood samples, there was no relationship between blood sulphide and protein intake despite an increase in faecal sulphide. Because of its toxicity, the sulphide detected in blood is present at low levels, possibly bound to protein [18], whilst the majority is excreted in faeces or is detoxified by colonocytes or red blood cells by thiolmethyl transferase [17]. The method described here is simpler and more accurate over a wider range of values than those previously published for the detection of sulphide in intestinal contents. Limitations previously encountered in some other methods, such as sample turbidity as seen in colorimetric determinations, were avoided by trapping anionic sulphide in a stable, non-particulate solution. This, coupled with highly selective and sensitive DC amperometric detection, achieved greater accuracy and user accessibility than previously reported for biological samples, and should allow the exploration of factors determining sulphide metabolism in humans.
Acknowledgements This work was supported by a grant from the Ministry of Agriculture Fisheries and Food.
References [1] Evans CL. The toxicity of hydrogen sulphide and other sulphides. Quart J Exp Physiol 1967;52:231–48. [2] Goodwin LR, Francom D, Dieken FP et al. Determination of sulfide in brain tissue by gas dialysis / ion chromatography: Postmortem studies and two case reports. J Anal Toxicol 1989;13:105–9. [3] Gibson GR, Macfarlane GT, Cummings JH. Sulphate-reducing bacteria and hydrogen metabolism in the human large intestine. Gut 1993;34:437–9. [4] Magee EAM, Cummings JH. Fecal sulfide levels generated in response to meat intake. Gastroenterology 1997;112:A891.
C. J. Richardson et al. / Clinica Chimica Acta 293 (2000) 115 – 125
125
[5] Roediger WEW, Duncan A, Kapaniris O, Millard S. Sulphide impairment of substrate oxidation in rat colonocytes: A biochemical basis for ulcerative colitis? Clin Sci 1993;85:623–7. [6] Roediger WEW, Duncan A, Kapaniris O, Millard S. Reducing sulphur compounds of the colon impair colonocyte nutrition: Implications for ulcerative colitis. Gastroenterology 1993;104:802–9. [7] Aslam M, Batten JJ, Florin THJ, Sidebotham RL, Baron JR. Hydrogen sulphide induced damage to the colonic mucosal barrier in the rat. Gut 1992;33(2):569. [8] Christl SU, Eisner RD, Dusel G, Kasper H, Scheppach W. Antagonistic effects of sulfide and butyrate on proliferation of colonic mucosa: a potential role for these 17 agents in the pathogenesis of ulcerative colitis. Dig Dis Sci 1996;41:2477–81. [9] Elson CO, Sartour RB, Tennyson GS, Riddell RH. Experimental models of inflammatory bowel disease. Gastroenterology 1995;109:1344–67. [10] Onderdonk AB, Bartlett MD. Bacteriological studies of experimental ulcerative colitis. Am J Clin Nutr 1979;32:258–65. [11] Savage JC, Gould DH. Determination of sulfide in brain tissue and rumen fluid by ion-interaction reversed-phase high-performance liquid chromatography. J Chromatogr 1990;526(2):540–5. [12] Goodwin LR, Francom D, Alessandro U, Dieken FP. Determination of trace sulfides in turbid waters by gas dialysis / ion chromatography. Anal Chem 1988;60:216–9. [13] Hughes R, O’Neill IK, Pollock JRA, Cummings JH, Bingham S. Dose response effect with dietary meat on faecal N-Nitroso compound levels in humans. Gastrointest Oncol 1997;112:A581. [14] Rocklin RD, Johnson EL. Determination of cyanide, sulfide, iodide and bromide by ion chromatography with electrochemical detection. Anal Chem 1983;55:4–7. [15] Suarez FL, Fume JK, Springfield JR, Levitt MD. Bismuth subsalicylate markedly decreases hydrogen sulfide release in the human colon. Gastroenterology 1998;114:G1711. [16] MacAnally BR, Lowry WT, Oliver RD, Garriott JC. Determination of inorganic sulfide and cyanide in blood using specific ion electrodes: Application to the investigation of hydrogen sulfide and cyanide poisoning. J Anal Toxicol 1979;3:111–4. [17] Roediger WEW, Moore J, Babidge W. Colonic sulfide in pathogenesis and treatment of ulcerative colitis. Dig Dis Sci 1997;42:1571–9. [18] Togawa T, Owaga M, Nawata M, Ogasawara Y, Kawanabe K, Tanabe S. High performance liquid chromatographic determination of bound sulfide and sulfite and thiosulfate at their low levels in human serum by precolumn fluorescence derivatization with monobromobimane. Chem Pharm Bull 1992;40:3000–4. [19] Magee EAM, Cummings JR. The contribution of dietary protein from meat to faecal sulphide excretion. Gut 1997;40:A27. [20] Pitcher MCL, Beatty ER, Cummings JH. The contribution of sulfate-reducing bacteria and 5-ASA to fecal sulfide in patients with ulcerative colitis. Gut 1999: in press.
A new method for the determination of sulphide in gastrointestinal contents and whole blood by microdistillation and ion chromatography Caroline J. Richardson, Elizabeth A.M. Magee, John H. Cummings* MRC Dunn Clinical Nutrition Centre, Cambridge, UK Received 1 March 1999; received in revised form 15 November 1999; accepted 17 November 1999
Abstract Hydrogen sulphide is produced in the human large intestine by the bacterial reduction of dietary inorganic sulphate and sulphite and by fermentation of sulphur amino acids. Sulphide may damage the colonic epithelium and has been implicated in the pathogenesis of ulcerative colitis. The accurate measurement of sulphide in biological samples, particularly in gut contents is difficult due to the volatile nature of the compound, and the viscosity and turbidity of the samples. Here we describe a method for the determination of sulphide in gut contents and whole blood which overcomes these problems. Initially, samples are treated with zinc acetate to trap sulphide. A microdistillation pretreatment is then used, which releases sulphide from its stable, stored state, coupled to ion chromatography with electrochemical detection. The limit of detection of the method was determined as 2.5 mmol / l, which enabled sulphide levels in gut contents and whole blood samples obtained from humans to be accurately determined. A preliminary investigation in healthy human subjects showed blood sulphide ranged from 10 to 100 mmol / l. Whole blood sulphide did not change significantly when increasing amounts of protein from meat were fed. However, faecal sulphide did show a significant increase from 164 to 754 nmol / g in four subjects fed diets which contained 60 and 420 g meat. 2000 Elsevier Science B.V. All rights reserved.
1. Introduction Hydrogen sulphide is an abundant by-product of both naturally occurring and *Corresponding author. Present address: Molecular and Cellular Pathology, Ninewells Hospital and Medical School, Dundee DDl 9SY, UK. 0009-8981 / 00 / $ – see front matter 2000 Elsevier Science B.V. All rights reserved. PII: S0009-8981( 99 )00245-4
116
C. J. Richardson et al. / Clinica Chimica Acta 293 (2000) 115 – 125
industrial processes. It has a high level of toxicity, comparable to that of hydrogen cyanide [1] and acts by the inhibition of cytochrome oxidase and thus arrests aerobic respiration, especially in nervous tissue. Several cases of fatal poisoning from H 2 S gas have been reported, and this has led to the development of a diverse range of detection methods [2]. In the human large intestine, hydrogen sulphide is produced as a product of bacterial metabolism by two known pathways. Firstly by the reduction of sulphate by sulphate-reducing bacteria (SRB) [3] and secondly by the fermentation of sulphur-containing amino acids, cysteine and methionine [4]. The presence of this toxic metabolite in the colon may be of importance in disease — in particular it has been implicated in the pathogenesis of ulcerative colitis. In vitro in both rat and human colonocytes, Roediger et al. [5,6] have shown that HS 2 at 2 mmol / l inhibits butyrate oxidation. This biochemical lesion is characteristic of the defect seen in ulcerative colitis and is due to persulphide formation which blocks butyryl CoA dehydrogenase. In perfusion studies of the rat colon, HS 2 at 0.2 mmol / l produces mucosal ulceration, goblet cell loss, apoptosis and distortion of the crypt architecture [7]. Using human colon tissue, Christl et al. [8] have shown that HS 2 at 1 mmol / l significantly increases cell proliferation rates. Furthermore, early studies of ulcerative colitis [9] showed that reliable animal models can be devised by feeding sulphated polysaccharides such as dextran sulphate and carageenan. These polysaccharides deliver sulphates directly to the colon. Bacteria are essential for this model [10]. An accurate and sensitive method for the detection of sulphide in gut contents is therefore important. The determination of volatile compounds such as sulphide in biological samples presents several obstacles. Sulphide is readily oxidised in air and in addition adheres to glass and rubber. Samples must therefore be prepared and stored in order to limit the loss of H 2 S gas. This is usually achieved by trapping the sulphide as stable zinc sulphide precipitate immediately a sample is obtained. The formation of the metal precipitate then poses problems in itself. Turbidity of the samples (in the case of gut contents) causes interference with colorimetrically determined assays such as the methylene blue, and these assays also lack sensitivity [11]. Direct analysis of samples where sulphide has been trapped as a metal complex by methods such as ion chromatography is unrealistic, since the complex formed is strong and therefore produces poor peak shapes and poor reproducibility as the complex slowly dissociates. It is therefore necessary to release the sulphide from its complex e.g. by the addition of acid and trap it in another solution. This has previously been achieved in turbid water and brain tissue samples using a gas dialysis procedure [2,12]. However, this method employs highly specialised and expensive apparatus and specially constructed gas dialysis blocks. Here, we present a microdistillation procedure in which H 2 S is first trapped as
C. J. Richardson et al. / Clinica Chimica Acta 293 (2000) 115 – 125
117
stable sulphide complexes and then distilled into a sodium hydroxide solution and analysed by ion chromatography. Preliminary results of sulphide in blood and faeces from human feeding studies are also given.
2. Method
2.1. Volunteers and diets Fresh faecal samples were collected from 15 free-living healthy volunteers. Blood was collected from six similarly free-living volunteers. Four healthy subjects, who were taking part in a controlled feeding study in the metabolic suite at the Dunn Clinical Nutrition Centre in which different amounts of meat were fed to observe the effects on faecal excretion of N-Nitroso compounds, also gave a blood sample. This was taken fasting on the final day of a 15-day diet period during which the subjects were eating either a no meat diet, or one containing 240 or 420 g meat [13]. Similarly four volunteers taking part in a controlled diet study of the effect of meat and sulphur additives in the diet on sulphur metabolism in the gut gave a faecal sample. The diets contained either 60 g meat or 420 g meat with sulphur additives.
2.2. Preparation and storage of samples All solutions used, with the exception of sodium hydroxide and eluent solutions, were deoxygenated by autoclaving at 1208C, 15 p.s.i, for 20 min, then topping up the solutions to the brim while still hot and leaving to cool tightly sealed. A 10% (w / v) faecal slurry was made by homogenising the freshly voided faecal sample in 0.1 mol / l potassium phosphate buffer pH 7.4 for 20 min in a stomacher. (Colworth Stomacher 400). The homogenate was then filtered through a 500-mm sieve to remove particulate matter, and 4 ml of filtrate was added to 1 ml of deoxygenated 120-g / l zinc acetate in a 7-ml bijou bottle. This traps the sulphide as zinc sulphide precipitate. Validation experiments using portions of a freshly passed faecal specimen showed that there was no difference in sulphide recovery if the slurry with zinc acetate was prepared as stated above or by homogenising the faecal specimen in a mixed trapping solution of anoxic phosphate buffer and zinc acetate prior to sieving. A single faecal slurry was divided into multiple aliquots and spiked and unspiked samples were stored in 7-ml bijou bottles with aluminium caps with rubber seals for up to 6 months at 48C without loss of sulphide.
118
C. J. Richardson et al. / Clinica Chimica Acta 293 (2000) 115 – 125
Whole blood was collected using 5-Monovette EDTA KE blood tubes (Sarstedt, Leicester, UK) and stored immediately at 2 208C.
2.3. Microdistillation procedure The microdistillation apparatus is shown in Fig. 1. After several prototypes had been tried, the trap units were manufactured by Soham Scientific (Ely, UK). The traps consist of two glass tubes, one for the test sample which has an injection port into which acid is injected to release hydrogen sulphide. The other tube is used to trap the hydrogen sulphide as anionic sulphide in 1 mol / l of sodium hydroxide solution. The two tubes are joined by a head unit which enables H 2 S gas evolved to be carried into the trapping solution. The design also allows the introduction of inert carrier gas directly into the sample tube, so mixing the sample throughout the procedure. The deoxygenated trapping solution was prepared by adding 40 g of sodium hydroxide pellets (BDH) to 1 l of deoxygenated water on the day of use. One ml of the trapping solution was pipetted into the trapping side of the unit and 1 ml of well-mixed sample was pipetted into the sample side of the unit. For blood, serum and other proteinaceous samples, 50 ml of a 10% (v / v) solution of antifoam SE-35 (Sigma) was added to 450 ml of 120 g / l zinc acetate and 500 ml of sample. The
Fig. 1. Microdistillation apparatus.
C. J. Richardson et al. / Clinica Chimica Acta 293 (2000) 115 – 125
119
head unit was then secured in position by sealing the joints with high vacuum silicone grease (BDH) and keck clips (Quickfit). In order to speed up the trapping process and improve recoveries, the sample side of the units were placed into a water bath at 508C, leaving the trapping side at room temperature. Increasing the temperature at which the evolution of HS 2 occurs greatly improved recoveries. However, increasing the temperature above 508C resulted in poor reproducibility and recoveries due to sample evaporation and condensation problems. Gas lines were attached to each unit and oxygen free nitrogen was bubbled through the traps at 200 cm 3 / min. After 5 min, 1 ml of concentrated HCl (BDH, Lutterworth, UK) was added to the sample via the injection port using a 21 g A2 needle (Microlance) and 1-ml syringe. The units were then left for 3 h at 508C. During assay validation recoveries of greater than 85% were achieved in 1.5 h under these conditions, but in order to increase recoveries consistently to above 85% and ensure day to day reproducibility, the time of the procedure was increased to 3 h. A volume of 300 ml of trapping solution was then transferred to a 0.5-ml polyvial (Dionex P/ N 38142) and 300 ml eluent added.
2.4. Detection of sulphide by ion chromatography Eluent was prepared by dissolving 0.6753 g oxalic acid (Sigma) in 984 ml of ultra pure water, purging with helium to degas, then adding 16 ml of 50% (w / v) solution sodium hydroxide. A Dionex DX-500 chromatography system, with a Carbopac PA-1 column 4 3 250 mm (Dionex PIN 35391) and Carbopac PA guard column 3 3 25 mm (Dionex P/ N 37141) were used with a 50-ml sample loop and 1 ml / min eluent flow-rate. A silver working electrode electrochemical cell was used (Dionex P/ N 44110), set at 0.00 V applied potential and output range 1 nA. It was found that the continuous injection of sulphide produced an oxidised precipitate Ag 2 S on the surface of the electrode. This precipitate does not effect the sensitivity of the electrode [14] though to prevent excessive build-up, the electrode surface was cleaned before each day of use, usually 80 samples, by gently rubbing with a clean pencil eraser and rinsing with ultra pure water ( . 18 MV)). The system is then allowed to equilibrate after which six injections of a 50 mmol / l sulphide standard solution are injected and reproducibility checked, as reported previously by Rocklin and Johnson [14]. Reproducibility was best if the sample vials were filled with 300 ml of sample and 300 ml of eluent, a total volume of 600 ml which was the full capacity of the vials. Calibration standards of 50, 20 and 10 mmol / l were made up in 1 mol / l sodium hydroxide on the day of analysis. Run time for a batch of 80 samples was 10 h.
120
C. J. Richardson et al. / Clinica Chimica Acta 293 (2000) 115 – 125
3. Results
3.1. Validation of ion chromatography detection The current response, due to the oxidation or reduction of solutes that occur across the electrode surface, which is measured by electrochemical detectors, is dependent upon the potential applied to the electrode (Eapp ). In this system the optimum Eapp for a 20 mmol / l ammonium sulphide standard was determined as 0 V. A calibration curve is constructed for each run. The linear range of the calibration curve was found to be 10 mmol / l to 1000 mmol / l. The limit of detection was determined as 2.5 mmol / l. Repetitive injections of a 20 mmol / l ammonium sulphide standard over 7 h gave a relative standard deviation of 3.9%. It was therefore concluded that sulphide stored in eluent was stable enough to be left in polyvials overnight in the autosampler, and this enabled us to run much larger batches of samples.
3.2. Validation of the microdistillation procedure Table 1 details the recovery of sulphide from faecal slurries which had been stored in zinc acetate and then spiked with varying concentrations of ammonium sulphide standard before the microdistillation procedure. Using this method, recoveries of 0.25, 0.5, 0.75, 1.0 and 1.5 mmol / l spikes were 98.365.3%, 101.766.9%, 94.0%, 95.563.5% and 98% respectively. Recoveries for the methylene blue assay in our laboratory ranged from 77.1 to 91.5%. Recovery experiments for the microdistillation method were also performed on the whole Table 1 Recovery of ammonium sulphide from spiked faecal slurries Sample concentration (nmol / g wet weight faeces) 6S.D.
(NH 4 ) 2 S spike concentration (nmol / g wet weight faeces)
Recovery (%)
110610 18064 15064 12606100 60065 15067 6061 72063 72062 52064 35065
1500 1000 1000 750 500 500 500 250 250 250 250
98 99 92 94 108 105 92 97 96 107 93
C. J. Richardson et al. / Clinica Chimica Acta 293 (2000) 115 – 125
121
Table 2 Recovery of ammonium sulphide from spiked whole blood samples in six subjects Blood sulphide concentration (mmol / l)6S.D
Concentration of (NH 4 ) 2 S spike (mmol / l)
Recovery (%)
2.160.03 2.660.10 3.060.10 2.560.05 5.160.59 6.0460.16
5.0 5.0 5.0 10.0 10.0 10.0
102 98 100 91 79 81
blood samples (Table 2) and were 10061.63% and 83.765.25% for the 5 and 10 mmol / l spike. Because hydrogen sulphide is such a volatile compound, it was also necessary to determine losses incurred throughout the entire procedure. This was achieved by adding a series of spikes ranging from 500 to 1000 mmol / l ammonium sulphide to freshly voided faecal samples before homogenisation. Recoveries ranged from 85.5 to 91.7%, determined using the microdistillation / IC procedure. We then compared faecal sample results obtained by this method and those obtained in parallel, on the same sample, by the methylene blue method (Table 3). A direct comparison of results obtained by the methylene blue assay and the Table 3 Comparison of sulphide values obtained from faecal slurries by the microdistillation / ion chromatography and the methylene blue assay Sulphide concentration determined by methylene blue assay (nmol / g wet weight faeces)
Sulphide concentration determined by microdistillation / IC (nmol / g wet weight faeces)
100 180 260 270 300 370 470 470 550 590 610 640 680 1700 1880
180 190 460 260 370 480 430 460 550 620 640 640 680 1200 1350
122
C. J. Richardson et al. / Clinica Chimica Acta 293 (2000) 115 – 125
microdistillation / ion chromatography method compared well between the two assays at most sample concentrations (r 2 5 0.91) although there appears to be overestimation of sulphide at concentrations over 1000 nmol / g wet weight with the methylene blue method. Fig. 2 shows the methylene blue calibration curve at concentrations above 5 mmol / l. This equates to an original sample concentration of 1000 nmol / g wet weight. Whole blood sulphide in four volunteers living on controlled diets with
Fig. 2. Methylene blue calibration curve (r 2 5 1.00).
C. J. Richardson et al. / Clinica Chimica Acta 293 (2000) 115 – 125
123
differing levels of meat intake were (mmol / l6S.E.M.) meat free 42 (15); 240 g of meat 47 (12); and 420 g of meat 25 (0.8). Faecal sulphides from a similar feeding study (nmol / g wet weight6S.E.M.) were 60 g of meat 164 (19) and 420 g of meat with sulphur additives 754 (140) P , 0.05.
4. Discussion Sulphide measurement in biological samples is difficult because of its volatility, susceptibility to oxidation, adsorption to glass and rubber and binding to organic molecules [15]. Because of this elaborate methods have been developed. The present method measures free sulphide and is simple and robust. Adsorption to glass and rubber is avoided by ensuring that samples do not come into contact with either substance prior to precipitation with zinc acetate. The addition of zinc acetate forms a very strong, stable complex preventing evolution of hydrogen sulphide. Overall recoveries from spiked fresh samples were less than 100%, but we believe that this is due to loss of H 2 S incurred during the initial homogenisation and filtration steps of sample preparation. It is therefore advisable that the time taken to prepare samples is kept to a minimum. Previously the detection of sulphide in diverse samples such as turbid and waste waters, brain tissue and serum have used a range of methods. These published methods have included the use of ion specific electrodes coupled to the Conway plate method [16], pre-column derivatisation of sulphide (based on the methylene blue assay [17] or monobromoamine derivatisation [18]) with detection by HPLC, and the use of gas dialysis with ion chromatography [2,12]. Ion specific electrodes have been reported to measure sulphide concentrations at 10 ng / ml (0.29 mmol / l), but these electrodes often give slow responses. They cannot be used directly on many biological samples because cyanide and sulphide can bind to components in the matrix. Attempts to separate cyanide and sulphide from the matrix using Conway plates have been documented, but no recovery data were published and we found poor recoveries of ammonium sulphide (0–60%) and irreproducible results from gut contents using microdiffusion plates. The use of pre-column derivatisation of sulphide by monobromoamine with HPLC separation and fluorescence detection is very sensitive (40 nmol / l) but requires very specialised equipment and it is also very labour intensive, unlike the microdistillation procedure. Faecal sulphides in a group of 15 volunteers ranged from 110 to 720 nmol / g wet weight. Feeding studies in healthy volunteers, which we have undertaken previously, have shown that faecal sulphide, measured by the methylene blue technique, can be manipulated by intake of protein from meat [19]. In these studies faecal sulphide concentrations ranged from 220 to 2800 nmol / g wet weight faeces. Other studies using the methylene blue assay have shown median
124
C. J. Richardson et al. / Clinica Chimica Acta 293 (2000) 115 – 125
(interquartile range) faecal sulphides of 250 (120–450) nmol / g in healthy controls and 680 (320–1070) nmol / g in untreated ulcerative colitis patients [20]. Using the new method we have shown in a preliminary analysis from an ongoing feeding study of healthy subjects that faecal sulphide increases significantly from 160 to 750 nmol / g when meat intake is increased, along with food containing sulphur additives as preservatives. Total protein intake in this study increased from 64 to 165 g per day. The lower detection limit of the new procedure enabled us to investigate sulphide levels in blood samples taken from volunteers on different protein diets. The data obtained showed that although sulphide could be easily detected in blood samples, there was no relationship between blood sulphide and protein intake despite an increase in faecal sulphide. Because of its toxicity, the sulphide detected in blood is present at low levels, possibly bound to protein [18], whilst the majority is excreted in faeces or is detoxified by colonocytes or red blood cells by thiolmethyl transferase [17]. The method described here is simpler and more accurate over a wider range of values than those previously published for the detection of sulphide in intestinal contents. Limitations previously encountered in some other methods, such as sample turbidity as seen in colorimetric determinations, were avoided by trapping anionic sulphide in a stable, non-particulate solution. This, coupled with highly selective and sensitive DC amperometric detection, achieved greater accuracy and user accessibility than previously reported for biological samples, and should allow the exploration of factors determining sulphide metabolism in humans.
Acknowledgements This work was supported by a grant from the Ministry of Agriculture Fisheries and Food.
References [1] Evans CL. The toxicity of hydrogen sulphide and other sulphides. Quart J Exp Physiol 1967;52:231–48. [2] Goodwin LR, Francom D, Dieken FP et al. Determination of sulfide in brain tissue by gas dialysis / ion chromatography: Postmortem studies and two case reports. J Anal Toxicol 1989;13:105–9. [3] Gibson GR, Macfarlane GT, Cummings JH. Sulphate-reducing bacteria and hydrogen metabolism in the human large intestine. Gut 1993;34:437–9. [4] Magee EAM, Cummings JH. Fecal sulfide levels generated in response to meat intake. Gastroenterology 1997;112:A891.
C. J. Richardson et al. / Clinica Chimica Acta 293 (2000) 115 – 125
125
[5] Roediger WEW, Duncan A, Kapaniris O, Millard S. Sulphide impairment of substrate oxidation in rat colonocytes: A biochemical basis for ulcerative colitis? Clin Sci 1993;85:623–7. [6] Roediger WEW, Duncan A, Kapaniris O, Millard S. Reducing sulphur compounds of the colon impair colonocyte nutrition: Implications for ulcerative colitis. Gastroenterology 1993;104:802–9. [7] Aslam M, Batten JJ, Florin THJ, Sidebotham RL, Baron JR. Hydrogen sulphide induced damage to the colonic mucosal barrier in the rat. Gut 1992;33(2):569. [8] Christl SU, Eisner RD, Dusel G, Kasper H, Scheppach W. Antagonistic effects of sulfide and butyrate on proliferation of colonic mucosa: a potential role for these 17 agents in the pathogenesis of ulcerative colitis. Dig Dis Sci 1996;41:2477–81. [9] Elson CO, Sartour RB, Tennyson GS, Riddell RH. Experimental models of inflammatory bowel disease. Gastroenterology 1995;109:1344–67. [10] Onderdonk AB, Bartlett MD. Bacteriological studies of experimental ulcerative colitis. Am J Clin Nutr 1979;32:258–65. [11] Savage JC, Gould DH. Determination of sulfide in brain tissue and rumen fluid by ion-interaction reversed-phase high-performance liquid chromatography. J Chromatogr 1990;526(2):540–5. [12] Goodwin LR, Francom D, Alessandro U, Dieken FP. Determination of trace sulfides in turbid waters by gas dialysis / ion chromatography. Anal Chem 1988;60:216–9. [13] Hughes R, O’Neill IK, Pollock JRA, Cummings JH, Bingham S. Dose response effect with dietary meat on faecal N-Nitroso compound levels in humans. Gastrointest Oncol 1997;112:A581. [14] Rocklin RD, Johnson EL. Determination of cyanide, sulfide, iodide and bromide by ion chromatography with electrochemical detection. Anal Chem 1983;55:4–7. [15] Suarez FL, Fume JK, Springfield JR, Levitt MD. Bismuth subsalicylate markedly decreases hydrogen sulfide release in the human colon. Gastroenterology 1998;114:G1711. [16] MacAnally BR, Lowry WT, Oliver RD, Garriott JC. Determination of inorganic sulfide and cyanide in blood using specific ion electrodes: Application to the investigation of hydrogen sulfide and cyanide poisoning. J Anal Toxicol 1979;3:111–4. [17] Roediger WEW, Moore J, Babidge W. Colonic sulfide in pathogenesis and treatment of ulcerative colitis. Dig Dis Sci 1997;42:1571–9. [18] Togawa T, Owaga M, Nawata M, Ogasawara Y, Kawanabe K, Tanabe S. High performance liquid chromatographic determination of bound sulfide and sulfite and thiosulfate at their low levels in human serum by precolumn fluorescence derivatization with monobromobimane. Chem Pharm Bull 1992;40:3000–4. [19] Magee EAM, Cummings JR. The contribution of dietary protein from meat to faecal sulphide excretion. Gut 1997;40:A27. [20] Pitcher MCL, Beatty ER, Cummings JH. The contribution of sulfate-reducing bacteria and 5-ASA to fecal sulfide in patients with ulcerative colitis. Gut 1999: in press.