Anal Bioanal Chem DOI 10.1007/s00216-015-8533-5
RESEARCH PAPER
Rapid and sensitive LC–MS–MS determination of 2-mercaptobenzothiazole, a rubber additive, in human urine Wolfgang Gries & Katja Küpper & Gabriele Leng
Received: 26 September 2014 / Revised: 29 January 2015 / Accepted: 2 February 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract 2-Mercaptobenzothiazole (MBT) is one of the most important vulcanization accelerators in the industrial production of rubber, especially car tires. Given its wide use in household articles and industrial rubber products it has a high potential to migrate into the environment. Humans can be exposed by dermal, oral, or inhalative routes. Incorporated MBT is excreted in urine, mainly as conjugates to glucuronide, sulfate, and mercapturic acid. On the basis of these facts MBT has been selected as a substance of high interest in the large scale 10-year German project on human biomonitoring (HBM); a cooperation between the German Federal Ministry for the Environment (BMUB) and the German Chemical Industry Association (VCI) with the objective of developing new analytical methods for relevant chemicals. The presented method was developed to determine MBT in human urine to reliably investigate the internal human MBT dose. Total MBT is measured after enzymatic hydrolysis followed by application of high-pressure liquid chromatography tandem mass spectrometry (HPLC–MS–MS) in positive-electrosprayionization mode (ESI+) using isotope-dilution quantification. High sample throughput could be obtained by use of the column-switching technique. Optimization yielded an analytical method with a low and reproducible limit of detection (LOD) of 0.4 μg L−1 and a limit of quantification (LOQ) of 1 μg L−1, and low relative standard deviations in the range 1.6–5.8 %. A small biomonitoring study covering unexposed humans and occupationally exposed workers was performed to establish the feasibility and reliability of the method. MBT was found in only one urine sample from the unexposed humans, at a value of 10.8 μg MBT per liter, whereas it was
W. Gries (*) : K. Küpper : G. Leng Health Protection, Institute of Biomonitoring, Currenta GmbH & Co. OHG, 51368 Leverkusen, Germany e-mail: [email protected]
found in all samples from the tested workers at values of up to 6210 μg MBT per liter. Keywords 2-Mercaptobenzothiazole . MBT . Urine . Conjugates . Excretion . LC–MS–MS . Biomonitoring
Introduction 2-Mercaptobenzothiazole (MBT, benzothiazole-2-thiol, CAS 149-30-4) was introduced to rubber production in the late 1940s [1]. It has been marketed under different trade names, including Captax, Accelerator, Mertax, and Vulkacit Merkapto. MBT consists of slightly yellowish crystals with a melting point of approximately 182 °C [2] and exists in the two tautomeric structures 2-benzothiazolethione (NH-form) and 2-benzothiazolethiole (SH-form), in which the equilibrium is shifted to 2-benzothiazolethione (Fig. 1). MBT is potentially found in many everyday articles (e.g. tires, cables, rubber gloves, rubber bands, seals, shoes, and drilling and cutting oil) and in rubber household articles (e.g. toys, utility gloves, and swimwear) [2, 3]. An additional application is the use of its sodium and zinc salts as a fungicide, used in latex, oil paints, and textile fibers [4]. Moreover, MBT c a n b e l i b e r a t e d a s a de co m p o s i t i o n pr o d u c t o f 2-(thiocyanomethylthio)benzothiazole [5], a widely used biocide for leather, pulp, and paper products [6], and of other MBT-containing accelerators of the sulfeneamide group [6, 7]. Given its ubiquitous application, MBT has a high potential to migrate into the environment [8]. Important sources of migration include abrasion of car tires and improper disposal of waste articles containing MBT. Several studies describe the ubiquitous distribution of MBT in the environment, e.g. water or soil [9], and its biodegradation [10]. A moderate skin-sensitizing potential of MBT was observed in animal experiments. A sensitizing potential was also
W. Gries et al. NH
N C
C
SH
S
S
S
Fig. 1 Tautomer structures of 2-mercaptobenzothiazole
observed in humans, particularly individuals who came into contact with rubber for extended periods during their professional work (e.g. shoemakers) or persons working with protective gloves (e.g. cleaning staff). This is documented in several human studies, e.g. using the patch tests [11]. In some epidemiological studies a carcinogenic effect has been discussed, but the results were inconclusive because mixed exposures with other chemicals (e.g. aromatic amines) could not be excluded [12–20]. MBT is not mutagenic or carcinogenic according to a comprehensive report by the German Federal Institute for Occupational Safety and Health. A reevaluation of human toxicity within the EU REACH Community Rolling Action Plan recently came to the conclusion that the available data are sufficient and appropriate to conclude that there is no need for a proposal for harmonized classification and labeling of 2-MBT [21]. From metabolism studies of rats it is known that more than 90 % of orally administered MBT is excreted in urine within 96 h. The predominant metabolites were MBT conjugated to glucuronide, sulfate, and mercapturic acid [5, 6]. Several publications in the field of toxicology or environmental distribution are available, but there are none in which the analytical conditions for MBT in human biomonitoring are tested for quantity, reproducibility, and robustness. Because of its wide application, the potential exposure of the general population, and the need for an analytical biomonitoring method, MBT was selected as a substance of high interest by the large-scale 10-year German project of human biomonitoring; a cooperation agreed between the German Federal Ministry for the Environment (BMUB) and German Chemical Industry Association (VCI). Therefore, our objective was to develop an analytical method for measuring MBT in urine, covering the wide concentration range expected both in environmental and occupational human-exposure scenarios.
purchased from Roche (Mannheim, Germany). All chemicals used were of the best available quality. The analytical standard 2-mercaptobenzothiazole (99 %) was obtained from Fluka, and the deuterated internal standard 2-mercaptobenzothiazole-(2,4,5,6-d4) was custom synthesized at IDM (Teltow, Germany). The identification and purity of the deuterated internal standard were confirmed by 1 H-NMR and 13C-NMR. The purity was ≥97 % and the isotopic purity was ≥98 %. Standard preparation For the preparation of the stock solution, 10 mg MBT was weighed exactly into a 10 mL glass volumetric flask and diluted to volume with acetonitrile (1000 mg L−1). From this stock solution a starting solution was prepared by diluting 100 μL in a 10 mL glass volumetric flask filled to the mark with acetonitrile. This starting solution (10 mg L−1) was gradually diluted for the preparation of the working standards, to achieve standard concentrations of 1000 μg L−1, 100 μg L−1, and 10 μg L−1. The internal-standard stock solution was prepared by diluting 10 mg MBT in a 10 mL volumetric flask with acetonitrile (1000 mg L−1). Spiking solution was made by diluting 100 μL to the mark in a 10 mL volumetric flask (10 mg L−1) with acetonitrile. Sample preparation Spot urine samples were collected in 250 mL pre-cleaned glass bottles (discussed in chapter BApplication for human biomonitoring^). The samples were stored at approximately −20 °C before analysis. For the analytical determination the samples were homogenized and a 0.5 mL aliquot was transferred into a 2 mL pre-cleaned crimp vial. Then 1 mL 1 mol L−1 ammonium acetate buffer pH 6.5, 10 μL MBT-d4 (10 mg L−1), and 5 μL β-glucuronidase/arylsulfatase were added. The vials were closed by a Teflon®-sealed crimp cap, homogenized by hand shaking, and warmed overnight at 37 °C in an incubator. Subsequently, samples were centrifuged at 2200g and placed in the HPLC autosampler.
Materials and methods
Calibration procedure and quantification
Chemicals
Calibration was performed by spiking 0.5 mL urine aliquots at 10 calibration points with concentrations ranging from 1 μg L−1 to 10,000 μg L−1 (0.6 nmol L−1–60 μmol L−1), as described in the chapters BStandard preparation^ and BSample collection^. The prepared calibration samples were determined by LC–MS–MS. Quadratic calibration curves with a weighting factor 1/x were obtained by plotting the quotient of the peak area from the target analyte and the corresponding
Acetonitrile, glacial acid, and hydrochloric acid 37 % were purchased from Merck (Darmstadt, Germany). Water was obtained from a millipore water-cleaning system, formic acid and ammonium acetate were purchased from Fluka (Taufkirchen, Germany), and β-glucuronidase/arylsulfatase Helix pomatia (5.5 u mL−1 and 2.6 u mL−1 at 38 °C) was
LC-MS-MS determination of MBT in human urine
deuterated internal standard against the standard concentrations. Quality control and validation Because there is no control material available, it was prepared in the laboratory using spiked combined urine samples at different concentrations (10, 100, and 1000 μg L−1 or 0.06, 0.6, and 6.0 μmol L−1). These control samples were stored in a refrigerator at −20 °C, and two samples of each concentration were analyzed during the analysis sequences on five different days to determine inter-day precision data. The intra-day precision data was obtained by analyzing spiked urine samples at the concentrations described above. These samples were analyzed 10 times in a row, and all samples were quantified against the calculated calibration curve. Moreover, the robustness of the analytical procedure was tested using 10 individual urine samples (creatinine: 0.5–2.8 g L−1) without detectable MBT spiked at two concentrations (10 and 100 μg L−1 or 0.06 and 0.6 μmol L−1). High-pressure liquid chromatography tandem mass spectrometry High-pressure-liquid-chromatography (HPLC) analysis was processed using a Waters Alliance 2695 HPLC system equipped with a quaternary pump, an autosampler, a degasser, and a special designed connection of columns and valves by tpieces (Waters, Eschborn, Germany). The eluents used consisted of a ternary gradient system with Milli-Q water (eluent A), 1 % formic acid in water (eluent B), and acetonitrile (eluent C) (Table 1). Separation of the target compound was performed by column switching. The two columns used were tempered at 30 °C, and 100 μL sample was injected. The first column (Waters Oasis HLB 2.1×20 mm, 25 μm) was used for sample clean-up and enrichment of the target analyte at a constant flow of 0.2 mL min−1. After 3.5 min the first column was connected to the second column (Zorbax Eclipse XDBC8 column (Agilent, 4.6×50 mm, 5 μm) by use of a switching
valve and the separated analyte was eluted by back flush onto the second column. Mass fragmentation and mass detection were performed using a Waters Quattro Ultima triple quadrupole in positiveelectrospray-ionization mode (ESI+), with a capillary voltage of 3.5 kV and a cone voltage of 120 V. Nitrogen was used as cone gas (100 L h−1) and desolvation gas (300 °C, flow 636 L h−1). The other source conditions were adjusted to Hex (1) 0 V, Aperture 0 V, Hex (2) 0 V, and source temperature 150 °C. Argon was used as collision gas, and the settings for the analyzer were: LM (1) resolution 15, HM (1) resolution 15, ion energy (1) 2, entrance −2, collision 23, exit 2, LM (2) resolution 15, HM (2) resolution 15, ion energy (2) 2, and multiplier 650. All instrument settings described were optimized by manual tuning with regard to the highest peak intensity of parent and resulting daughter fragment (Table 2). Instrument processing for HPLC and MS–MS detection and data handling were performed using the instrument’s software Mass Lynx 4.1. Application for human biomonitoring To reveal the applicability of the method for quantifying MBT in urine, two different sample groups were selected. We investigated five individuals working at an MBT production plant. Four workers had direct contact with MBT during their work activities, and one was working in the plant’s administration; he was the control. During the consultation hour the physician explained why biomonitoring is recommended and that the urine samples would only be used for the verification of the described analytical method. The second group consisted of 40 randomized urine samples from persons not knowingly exposed to MBT. Their urine samples were measured to investigate whether there is a possible background exposure to MBT and to identify any interferences.
Results Tandem mass spectrometry
Table 1
HPLC eluent conditions
Time (min)
Eluent A, Milli-Q water (%)
Eluent B, 1 % formic acid (%)
Eluent C, acetonitrile (%)
0 4 5 8 8.5 12 13
70 70 0 0 70 70 70
10 10 10 10 10 10 10
20 20 90 90 20 20 20
For MBT and the deuterated internal standard MBT-d4 analyzed in ESI+ mode, the molecular ion [M + H]+ was used as Table 2 Analyte
Mass transition of MBT and MBT-d4 Parent mass (m/z)
Daughter mass (m/z)
Dwell Delay Retention (s) (s) time (min)
MBT 167.93 135.09/124.15 0.2 MBT-d4 (ISTD) 171.95 139.11/128.17 0.2
0.1 0.1
13.34 13.31
W. Gries et al.
precursor followed by specific fragmentation and mass transition of the product ion. MBT had a mass transition of m/z 167.93 to m/z 135.09 (loss of SH, Quantifier) and m/z 124.15 (loss of CSH, Qualifier). A similar mass fragmentation was determined for MBT-d4, with m/z 171.95 to m/z 139.11 (Quantifier) and m/z 128.17 (Qualifier). Reliability of the method In all experiments performed, the coefficients of determination of the calibration curves were higher than r2 =0.998. A quantification range of 1 μg L−1 to 10,000 μg L−1 was used for the occupational samples, whereas a lower range (1– 1000 μg L−1) was used for the samples from the general population. As described in the BQuality control and validation^ chapter, several quality-control samples were analyzed to determine the reliability of the analytical method. The relative standard deviation was in the range 1.6–2.3 % for the intra-day precision and 3.4–5.8 % for the inter-day precision (Table 3). Recovery experiments with 10 different individual urine samples (creatinine 0.5–2.8 g L−1) spiked with 10 μg L−1 (recovery 84 %, RSD 11.2 %) and 100 μg L−1 (recovery 95 %, RSD 2.5 %) MBT were also performed to establish the robustness and reproducibility of the method. The reproducibility (RSD 5.1 %, n=20) of the enzymatic release of MBT also confirms the reliability of the method. The detection limit was calculated by the calibration-curve method using the six lowest calibration points. The established LOD for MBT was 0.4 μg L−1 (0.0024 μmol L−1) in urine, and the corresponding LOQ was calculated to be 1 μg L−1 (0.006 μmol L−1) (Fig. 2). Improvement of the method and sources of error Having urine samples from persons in contact with MBT made it possible to compare several hydrolysis procedures. This investigation was done to get an overview of the efficiency of the liberation of conjugated MBT in real urine samples. For all these experiments the urine samples of the four
Table 3
exposed workers were used. Three different analytical steps were tested: 1. acid hydrolysis at 20 °C for 1 h; 2. acid hydrolysis at 100 °C for 2 h; and 3. enzymatic hydrolysis at 37 °C for 2 h It was found that the enzymatic-hydrolysis procedure obtained up to fourfold higher MBT levels than the experiments with acid hydrolysis at room temperature [5]. Moreover, in the samples which underwent acidic hydrolysis by application of a hydrolysis procedure with 250 μL 37 % hydrochloric acid in 1 mL urine at 100 °C, MBT could not be measured correctly. This is because of the high matrix level, which causes ion suppression during the mass transition despite the use of deuterated internal standard. Hence, in this case the internal standard was added after the hydrolysis procedure to avoid deuterium replacement. Regarding the internal standard, it was also observed that 2-mercaptobenzoxazole (MBO), mentioned as an alternative internal standard in literature [5], could not be used for this method. Although the chemical structure of MBO is similar to that of MBT, MBO has a different fragmentation in ESI+ mode (m/z 152 to m/z 134 by loss of HOH) and a slightly shorter retention time. These differences resulted in an incomplete compensation for the analytical deviations. This observation is in line with our experience that a labeled internal standard is the best choice for analytical determination by HPLC–MS–MS. Because of an impurity of approximately 1 % native MBT content in the deuterated internal standard, the determination in the lower quantification range of environmental samples was found to be complicated if concentrations of 100 μg L−1 internal standard were spiked. Therefore these samples were analyzed with a lower concentration of 10 μg L−1 MBT-d4, and 100 μg L−1 was used for the occupational samples. The hydrolysis step with β-glucuronidase/arylsulfatase and the necessary incubation time and best pH value were also investigated. A combined urine sample consisting of the four previously analyzed urine samples from the workers was analyzed twenty times in a row. It was found that the endpoint of
Quality-control data for intra-day and inter-day precision (n=10)
Analyte Intra-day precision
MBT
Inter-day precision
Conc. (μg L−1)
Recovery (%)
R.S.D. (%)
Accuracy range (%)
Conc. (μg L−1)
Recovery (%)
R.S.D. (%)
Accuracy range (%)
10 100 1000
87 93 90
2.3 1.9 1.6
85–91 90–96 89–94
10 100 1000
86 96 96
4.3 5.8 3.4
81–93 93–112 91–102
Conc., concentration
LC-MS-MS determination of MBT in human urine Fig. 2 Chromatogram of a urine sample spiked with 1 μg MBT per liter
150118_55 Sm (Mn, 3x3) 13.34;3024;11135
MRM of 4 Channels ES+ 171.95 > 139.11 (MBT d4 ) 1.12e4 Area, Height
%
100
0 11.50 12.00 12.50 150118_55 Sm (Mn, 3x3)
13.00
13.50
13.32;1877;6458
%
100
14.00 14.50 MRM of 4 Channels ES+ 171.95 > 128.17 (MBT d4 _Q) 6.88e3 Area, Height
0 11.50 12.00 12.50 150118_55 Sm (Mn, 3x3)
13.00
13.50
13.36;561;1853
%
100
14.00 14.50 MRM of 4 Channels ES+ 167.93 > 135.09 (MBT) 2.03e3 Area, Height
0 11.50 12.00 12.50 150118_55 Sm (Mn, 3x3)
13.00
13.50
13.30;230;743
%
100
14.00 14.50 MRM of 4 Channels ES+ 167.93 > 124.15 (MBT_Q) 1.06e3 Area, Height
0 11.50
Time 12.00
the enzymatic hydrolysis step was reached within 30 min at the optimum pH of 6.5, and the relative standard deviation of the twenty samples analyzed for reproducibility was 5.1 %. Poor recoveries were obtained during the method development when sodium acetate was used as the buffer for the enzymatic hydrolysis. After switching to ammonium acetate buffer the recovery increased significantly. This was probably caused by the high ion-bonding affinity of MBT to sodium, which is much higher than that to ammonium. Another characteristic of MBT is the instability of the analytical standards. We observed a decrease of 20 % during the first three months, and after six months only 50 % of the initial value could be detected, although the standards were diluted in acetonitrile
12.50
13.00
13.50
14.00
14.50
and stored in the dark in a refrigerator at 5 °C. We recommended renewing the standard solutions every month. Because we did not have enough experience regarding possible contamination sources of MBT, all urine samples were collected in glass bottles cleaned with methanol and dichloromethane before use. The same cleaning procedure was used for the autosampler vials. This precaution was adopted on the basis of the results for the adsorption activity of MBT obtained during the method-development process and the possible leaching of MBT residues from the dishwasher’s rubber seals or packaging material. Later in the method development we concluded that normal one-way PP or PE vessels could also be used.
W. Gries et al.
Results of biological monitoring
Table 5 Results for four individual urine samples from workers exposed to MBT
Several urine samples were analyzed to get an overview of possible exposure levels of MBT. The presented method was tested on urine samples from 40 individuals occupationally unexposed to MBT for the determination of possible background exposure. In only one of these samples, a low level of 10.8 μg L−1 (5.3 μg g−1 creatinine) MBT was detectable (Table 4). In the other 39 samples no MBT was found at or above the LOQ of 1 μg L−1. In contrast with this sample group, high levels of MBT were determined in samples collected from workers at an MBT plant after their work shift. The one worker who had no direct contact with MBT and whose urine was used as a control sample had an MBT level of only 2.5 μg L−1 (2.9 μg g−1 creatinine), whereas the urine samples of his four colleagues had levels ranging from 567 up to 6210 μg L−1 (973–3051 μg g−1 creatinine). In these samples free MBT (without hydrolysis) was also measured (Table 5). Free MBT was detected in the range
RESEARCH PAPER
Rapid and sensitive LC–MS–MS determination of 2-mercaptobenzothiazole, a rubber additive, in human urine Wolfgang Gries & Katja Küpper & Gabriele Leng
Received: 26 September 2014 / Revised: 29 January 2015 / Accepted: 2 February 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract 2-Mercaptobenzothiazole (MBT) is one of the most important vulcanization accelerators in the industrial production of rubber, especially car tires. Given its wide use in household articles and industrial rubber products it has a high potential to migrate into the environment. Humans can be exposed by dermal, oral, or inhalative routes. Incorporated MBT is excreted in urine, mainly as conjugates to glucuronide, sulfate, and mercapturic acid. On the basis of these facts MBT has been selected as a substance of high interest in the large scale 10-year German project on human biomonitoring (HBM); a cooperation between the German Federal Ministry for the Environment (BMUB) and the German Chemical Industry Association (VCI) with the objective of developing new analytical methods for relevant chemicals. The presented method was developed to determine MBT in human urine to reliably investigate the internal human MBT dose. Total MBT is measured after enzymatic hydrolysis followed by application of high-pressure liquid chromatography tandem mass spectrometry (HPLC–MS–MS) in positive-electrosprayionization mode (ESI+) using isotope-dilution quantification. High sample throughput could be obtained by use of the column-switching technique. Optimization yielded an analytical method with a low and reproducible limit of detection (LOD) of 0.4 μg L−1 and a limit of quantification (LOQ) of 1 μg L−1, and low relative standard deviations in the range 1.6–5.8 %. A small biomonitoring study covering unexposed humans and occupationally exposed workers was performed to establish the feasibility and reliability of the method. MBT was found in only one urine sample from the unexposed humans, at a value of 10.8 μg MBT per liter, whereas it was
W. Gries (*) : K. Küpper : G. Leng Health Protection, Institute of Biomonitoring, Currenta GmbH & Co. OHG, 51368 Leverkusen, Germany e-mail: [email protected]
found in all samples from the tested workers at values of up to 6210 μg MBT per liter. Keywords 2-Mercaptobenzothiazole . MBT . Urine . Conjugates . Excretion . LC–MS–MS . Biomonitoring
Introduction 2-Mercaptobenzothiazole (MBT, benzothiazole-2-thiol, CAS 149-30-4) was introduced to rubber production in the late 1940s [1]. It has been marketed under different trade names, including Captax, Accelerator, Mertax, and Vulkacit Merkapto. MBT consists of slightly yellowish crystals with a melting point of approximately 182 °C [2] and exists in the two tautomeric structures 2-benzothiazolethione (NH-form) and 2-benzothiazolethiole (SH-form), in which the equilibrium is shifted to 2-benzothiazolethione (Fig. 1). MBT is potentially found in many everyday articles (e.g. tires, cables, rubber gloves, rubber bands, seals, shoes, and drilling and cutting oil) and in rubber household articles (e.g. toys, utility gloves, and swimwear) [2, 3]. An additional application is the use of its sodium and zinc salts as a fungicide, used in latex, oil paints, and textile fibers [4]. Moreover, MBT c a n b e l i b e r a t e d a s a de co m p o s i t i o n pr o d u c t o f 2-(thiocyanomethylthio)benzothiazole [5], a widely used biocide for leather, pulp, and paper products [6], and of other MBT-containing accelerators of the sulfeneamide group [6, 7]. Given its ubiquitous application, MBT has a high potential to migrate into the environment [8]. Important sources of migration include abrasion of car tires and improper disposal of waste articles containing MBT. Several studies describe the ubiquitous distribution of MBT in the environment, e.g. water or soil [9], and its biodegradation [10]. A moderate skin-sensitizing potential of MBT was observed in animal experiments. A sensitizing potential was also
W. Gries et al. NH
N C
C
SH
S
S
S
Fig. 1 Tautomer structures of 2-mercaptobenzothiazole
observed in humans, particularly individuals who came into contact with rubber for extended periods during their professional work (e.g. shoemakers) or persons working with protective gloves (e.g. cleaning staff). This is documented in several human studies, e.g. using the patch tests [11]. In some epidemiological studies a carcinogenic effect has been discussed, but the results were inconclusive because mixed exposures with other chemicals (e.g. aromatic amines) could not be excluded [12–20]. MBT is not mutagenic or carcinogenic according to a comprehensive report by the German Federal Institute for Occupational Safety and Health. A reevaluation of human toxicity within the EU REACH Community Rolling Action Plan recently came to the conclusion that the available data are sufficient and appropriate to conclude that there is no need for a proposal for harmonized classification and labeling of 2-MBT [21]. From metabolism studies of rats it is known that more than 90 % of orally administered MBT is excreted in urine within 96 h. The predominant metabolites were MBT conjugated to glucuronide, sulfate, and mercapturic acid [5, 6]. Several publications in the field of toxicology or environmental distribution are available, but there are none in which the analytical conditions for MBT in human biomonitoring are tested for quantity, reproducibility, and robustness. Because of its wide application, the potential exposure of the general population, and the need for an analytical biomonitoring method, MBT was selected as a substance of high interest by the large-scale 10-year German project of human biomonitoring; a cooperation agreed between the German Federal Ministry for the Environment (BMUB) and German Chemical Industry Association (VCI). Therefore, our objective was to develop an analytical method for measuring MBT in urine, covering the wide concentration range expected both in environmental and occupational human-exposure scenarios.
purchased from Roche (Mannheim, Germany). All chemicals used were of the best available quality. The analytical standard 2-mercaptobenzothiazole (99 %) was obtained from Fluka, and the deuterated internal standard 2-mercaptobenzothiazole-(2,4,5,6-d4) was custom synthesized at IDM (Teltow, Germany). The identification and purity of the deuterated internal standard were confirmed by 1 H-NMR and 13C-NMR. The purity was ≥97 % and the isotopic purity was ≥98 %. Standard preparation For the preparation of the stock solution, 10 mg MBT was weighed exactly into a 10 mL glass volumetric flask and diluted to volume with acetonitrile (1000 mg L−1). From this stock solution a starting solution was prepared by diluting 100 μL in a 10 mL glass volumetric flask filled to the mark with acetonitrile. This starting solution (10 mg L−1) was gradually diluted for the preparation of the working standards, to achieve standard concentrations of 1000 μg L−1, 100 μg L−1, and 10 μg L−1. The internal-standard stock solution was prepared by diluting 10 mg MBT in a 10 mL volumetric flask with acetonitrile (1000 mg L−1). Spiking solution was made by diluting 100 μL to the mark in a 10 mL volumetric flask (10 mg L−1) with acetonitrile. Sample preparation Spot urine samples were collected in 250 mL pre-cleaned glass bottles (discussed in chapter BApplication for human biomonitoring^). The samples were stored at approximately −20 °C before analysis. For the analytical determination the samples were homogenized and a 0.5 mL aliquot was transferred into a 2 mL pre-cleaned crimp vial. Then 1 mL 1 mol L−1 ammonium acetate buffer pH 6.5, 10 μL MBT-d4 (10 mg L−1), and 5 μL β-glucuronidase/arylsulfatase were added. The vials were closed by a Teflon®-sealed crimp cap, homogenized by hand shaking, and warmed overnight at 37 °C in an incubator. Subsequently, samples were centrifuged at 2200g and placed in the HPLC autosampler.
Materials and methods
Calibration procedure and quantification
Chemicals
Calibration was performed by spiking 0.5 mL urine aliquots at 10 calibration points with concentrations ranging from 1 μg L−1 to 10,000 μg L−1 (0.6 nmol L−1–60 μmol L−1), as described in the chapters BStandard preparation^ and BSample collection^. The prepared calibration samples were determined by LC–MS–MS. Quadratic calibration curves with a weighting factor 1/x were obtained by plotting the quotient of the peak area from the target analyte and the corresponding
Acetonitrile, glacial acid, and hydrochloric acid 37 % were purchased from Merck (Darmstadt, Germany). Water was obtained from a millipore water-cleaning system, formic acid and ammonium acetate were purchased from Fluka (Taufkirchen, Germany), and β-glucuronidase/arylsulfatase Helix pomatia (5.5 u mL−1 and 2.6 u mL−1 at 38 °C) was
LC-MS-MS determination of MBT in human urine
deuterated internal standard against the standard concentrations. Quality control and validation Because there is no control material available, it was prepared in the laboratory using spiked combined urine samples at different concentrations (10, 100, and 1000 μg L−1 or 0.06, 0.6, and 6.0 μmol L−1). These control samples were stored in a refrigerator at −20 °C, and two samples of each concentration were analyzed during the analysis sequences on five different days to determine inter-day precision data. The intra-day precision data was obtained by analyzing spiked urine samples at the concentrations described above. These samples were analyzed 10 times in a row, and all samples were quantified against the calculated calibration curve. Moreover, the robustness of the analytical procedure was tested using 10 individual urine samples (creatinine: 0.5–2.8 g L−1) without detectable MBT spiked at two concentrations (10 and 100 μg L−1 or 0.06 and 0.6 μmol L−1). High-pressure liquid chromatography tandem mass spectrometry High-pressure-liquid-chromatography (HPLC) analysis was processed using a Waters Alliance 2695 HPLC system equipped with a quaternary pump, an autosampler, a degasser, and a special designed connection of columns and valves by tpieces (Waters, Eschborn, Germany). The eluents used consisted of a ternary gradient system with Milli-Q water (eluent A), 1 % formic acid in water (eluent B), and acetonitrile (eluent C) (Table 1). Separation of the target compound was performed by column switching. The two columns used were tempered at 30 °C, and 100 μL sample was injected. The first column (Waters Oasis HLB 2.1×20 mm, 25 μm) was used for sample clean-up and enrichment of the target analyte at a constant flow of 0.2 mL min−1. After 3.5 min the first column was connected to the second column (Zorbax Eclipse XDBC8 column (Agilent, 4.6×50 mm, 5 μm) by use of a switching
valve and the separated analyte was eluted by back flush onto the second column. Mass fragmentation and mass detection were performed using a Waters Quattro Ultima triple quadrupole in positiveelectrospray-ionization mode (ESI+), with a capillary voltage of 3.5 kV and a cone voltage of 120 V. Nitrogen was used as cone gas (100 L h−1) and desolvation gas (300 °C, flow 636 L h−1). The other source conditions were adjusted to Hex (1) 0 V, Aperture 0 V, Hex (2) 0 V, and source temperature 150 °C. Argon was used as collision gas, and the settings for the analyzer were: LM (1) resolution 15, HM (1) resolution 15, ion energy (1) 2, entrance −2, collision 23, exit 2, LM (2) resolution 15, HM (2) resolution 15, ion energy (2) 2, and multiplier 650. All instrument settings described were optimized by manual tuning with regard to the highest peak intensity of parent and resulting daughter fragment (Table 2). Instrument processing for HPLC and MS–MS detection and data handling were performed using the instrument’s software Mass Lynx 4.1. Application for human biomonitoring To reveal the applicability of the method for quantifying MBT in urine, two different sample groups were selected. We investigated five individuals working at an MBT production plant. Four workers had direct contact with MBT during their work activities, and one was working in the plant’s administration; he was the control. During the consultation hour the physician explained why biomonitoring is recommended and that the urine samples would only be used for the verification of the described analytical method. The second group consisted of 40 randomized urine samples from persons not knowingly exposed to MBT. Their urine samples were measured to investigate whether there is a possible background exposure to MBT and to identify any interferences.
Results Tandem mass spectrometry
Table 1
HPLC eluent conditions
Time (min)
Eluent A, Milli-Q water (%)
Eluent B, 1 % formic acid (%)
Eluent C, acetonitrile (%)
0 4 5 8 8.5 12 13
70 70 0 0 70 70 70
10 10 10 10 10 10 10
20 20 90 90 20 20 20
For MBT and the deuterated internal standard MBT-d4 analyzed in ESI+ mode, the molecular ion [M + H]+ was used as Table 2 Analyte
Mass transition of MBT and MBT-d4 Parent mass (m/z)
Daughter mass (m/z)
Dwell Delay Retention (s) (s) time (min)
MBT 167.93 135.09/124.15 0.2 MBT-d4 (ISTD) 171.95 139.11/128.17 0.2
0.1 0.1
13.34 13.31
W. Gries et al.
precursor followed by specific fragmentation and mass transition of the product ion. MBT had a mass transition of m/z 167.93 to m/z 135.09 (loss of SH, Quantifier) and m/z 124.15 (loss of CSH, Qualifier). A similar mass fragmentation was determined for MBT-d4, with m/z 171.95 to m/z 139.11 (Quantifier) and m/z 128.17 (Qualifier). Reliability of the method In all experiments performed, the coefficients of determination of the calibration curves were higher than r2 =0.998. A quantification range of 1 μg L−1 to 10,000 μg L−1 was used for the occupational samples, whereas a lower range (1– 1000 μg L−1) was used for the samples from the general population. As described in the BQuality control and validation^ chapter, several quality-control samples were analyzed to determine the reliability of the analytical method. The relative standard deviation was in the range 1.6–2.3 % for the intra-day precision and 3.4–5.8 % for the inter-day precision (Table 3). Recovery experiments with 10 different individual urine samples (creatinine 0.5–2.8 g L−1) spiked with 10 μg L−1 (recovery 84 %, RSD 11.2 %) and 100 μg L−1 (recovery 95 %, RSD 2.5 %) MBT were also performed to establish the robustness and reproducibility of the method. The reproducibility (RSD 5.1 %, n=20) of the enzymatic release of MBT also confirms the reliability of the method. The detection limit was calculated by the calibration-curve method using the six lowest calibration points. The established LOD for MBT was 0.4 μg L−1 (0.0024 μmol L−1) in urine, and the corresponding LOQ was calculated to be 1 μg L−1 (0.006 μmol L−1) (Fig. 2). Improvement of the method and sources of error Having urine samples from persons in contact with MBT made it possible to compare several hydrolysis procedures. This investigation was done to get an overview of the efficiency of the liberation of conjugated MBT in real urine samples. For all these experiments the urine samples of the four
Table 3
exposed workers were used. Three different analytical steps were tested: 1. acid hydrolysis at 20 °C for 1 h; 2. acid hydrolysis at 100 °C for 2 h; and 3. enzymatic hydrolysis at 37 °C for 2 h It was found that the enzymatic-hydrolysis procedure obtained up to fourfold higher MBT levels than the experiments with acid hydrolysis at room temperature [5]. Moreover, in the samples which underwent acidic hydrolysis by application of a hydrolysis procedure with 250 μL 37 % hydrochloric acid in 1 mL urine at 100 °C, MBT could not be measured correctly. This is because of the high matrix level, which causes ion suppression during the mass transition despite the use of deuterated internal standard. Hence, in this case the internal standard was added after the hydrolysis procedure to avoid deuterium replacement. Regarding the internal standard, it was also observed that 2-mercaptobenzoxazole (MBO), mentioned as an alternative internal standard in literature [5], could not be used for this method. Although the chemical structure of MBO is similar to that of MBT, MBO has a different fragmentation in ESI+ mode (m/z 152 to m/z 134 by loss of HOH) and a slightly shorter retention time. These differences resulted in an incomplete compensation for the analytical deviations. This observation is in line with our experience that a labeled internal standard is the best choice for analytical determination by HPLC–MS–MS. Because of an impurity of approximately 1 % native MBT content in the deuterated internal standard, the determination in the lower quantification range of environmental samples was found to be complicated if concentrations of 100 μg L−1 internal standard were spiked. Therefore these samples were analyzed with a lower concentration of 10 μg L−1 MBT-d4, and 100 μg L−1 was used for the occupational samples. The hydrolysis step with β-glucuronidase/arylsulfatase and the necessary incubation time and best pH value were also investigated. A combined urine sample consisting of the four previously analyzed urine samples from the workers was analyzed twenty times in a row. It was found that the endpoint of
Quality-control data for intra-day and inter-day precision (n=10)
Analyte Intra-day precision
MBT
Inter-day precision
Conc. (μg L−1)
Recovery (%)
R.S.D. (%)
Accuracy range (%)
Conc. (μg L−1)
Recovery (%)
R.S.D. (%)
Accuracy range (%)
10 100 1000
87 93 90
2.3 1.9 1.6
85–91 90–96 89–94
10 100 1000
86 96 96
4.3 5.8 3.4
81–93 93–112 91–102
Conc., concentration
LC-MS-MS determination of MBT in human urine Fig. 2 Chromatogram of a urine sample spiked with 1 μg MBT per liter
150118_55 Sm (Mn, 3x3) 13.34;3024;11135
MRM of 4 Channels ES+ 171.95 > 139.11 (MBT d4 ) 1.12e4 Area, Height
%
100
0 11.50 12.00 12.50 150118_55 Sm (Mn, 3x3)
13.00
13.50
13.32;1877;6458
%
100
14.00 14.50 MRM of 4 Channels ES+ 171.95 > 128.17 (MBT d4 _Q) 6.88e3 Area, Height
0 11.50 12.00 12.50 150118_55 Sm (Mn, 3x3)
13.00
13.50
13.36;561;1853
%
100
14.00 14.50 MRM of 4 Channels ES+ 167.93 > 135.09 (MBT) 2.03e3 Area, Height
0 11.50 12.00 12.50 150118_55 Sm (Mn, 3x3)
13.00
13.50
13.30;230;743
%
100
14.00 14.50 MRM of 4 Channels ES+ 167.93 > 124.15 (MBT_Q) 1.06e3 Area, Height
0 11.50
Time 12.00
the enzymatic hydrolysis step was reached within 30 min at the optimum pH of 6.5, and the relative standard deviation of the twenty samples analyzed for reproducibility was 5.1 %. Poor recoveries were obtained during the method development when sodium acetate was used as the buffer for the enzymatic hydrolysis. After switching to ammonium acetate buffer the recovery increased significantly. This was probably caused by the high ion-bonding affinity of MBT to sodium, which is much higher than that to ammonium. Another characteristic of MBT is the instability of the analytical standards. We observed a decrease of 20 % during the first three months, and after six months only 50 % of the initial value could be detected, although the standards were diluted in acetonitrile
12.50
13.00
13.50
14.00
14.50
and stored in the dark in a refrigerator at 5 °C. We recommended renewing the standard solutions every month. Because we did not have enough experience regarding possible contamination sources of MBT, all urine samples were collected in glass bottles cleaned with methanol and dichloromethane before use. The same cleaning procedure was used for the autosampler vials. This precaution was adopted on the basis of the results for the adsorption activity of MBT obtained during the method-development process and the possible leaching of MBT residues from the dishwasher’s rubber seals or packaging material. Later in the method development we concluded that normal one-way PP or PE vessels could also be used.
W. Gries et al.
Results of biological monitoring
Table 5 Results for four individual urine samples from workers exposed to MBT
Several urine samples were analyzed to get an overview of possible exposure levels of MBT. The presented method was tested on urine samples from 40 individuals occupationally unexposed to MBT for the determination of possible background exposure. In only one of these samples, a low level of 10.8 μg L−1 (5.3 μg g−1 creatinine) MBT was detectable (Table 4). In the other 39 samples no MBT was found at or above the LOQ of 1 μg L−1. In contrast with this sample group, high levels of MBT were determined in samples collected from workers at an MBT plant after their work shift. The one worker who had no direct contact with MBT and whose urine was used as a control sample had an MBT level of only 2.5 μg L−1 (2.9 μg g−1 creatinine), whereas the urine samples of his four colleagues had levels ranging from 567 up to 6210 μg L−1 (973–3051 μg g−1 creatinine). In these samples free MBT (without hydrolysis) was also measured (Table 5). Free MBT was detected in the range
