Original
Paper
Chin-Yu Wang, Ching-Te Huang, You-Zung Hsieh Department of AppliedChemistry, National Chiao Tung University, Hsinchu, Taiwan, R.O.C.
Determination of metabolites of benzene, toluene,
ethylbenzene, and xylene by b-cyclodextrin
modified capillary electrophoresis
A b-cyclodextrin modified capillary electrophoresis method for determining the meta-bolites of benzene, toluene, ethyl benzene, and xylene (BTEX) is described. Creati-nine and uric acid were also examined as reference compounds. By employing a 20 mM, pH 10.0 borax-NaOH buffer containing 7-mM of b-CD, ten analytes could be successfully separated within 10 min on a 50 lm capillary at an electric potential of 25 kV. The effects of pH and the concentration of b-CD on the separation of the ten analytes are discussed. Under optimized separation conditions, the relative standard deviations for the migration times of the analytes were less than 0.93% and the corre-lation coefficients of the analyte linear calibration graphs exceeded 0.998 in the range from 1 to 300 lg/mL. Analyses of BTEX metabolites in urine samples from both occu-pationally exposed and non-exposed subjects are shown herein. BTEX metabolites were determined in the urine sample of a gas station attendant.
Key Words: BTEX metabolites; b-Cyclodextrin modified capillary electrophoresis;
Received: December 23, 2001; revised: August 22, 2002; accepted: September 12, 2002
1 Introduction
Benzene, toluene, ethylbenzene, and o-, m-, and p-xylenes (BTEX) are vital ingredients of commercial and industrial chemicals. They are widely used in fuels; for example, as primary components of motor vehicle gaso-line, as well as solvents, and starting products in a variety of chemical syntheses. These applications have rendered BTEX ubiquitous in the environment. In addition, these volatile organic compounds are highly toxic. They are easily absorbed via the lungs and the skin may also be a significant absorption route [1].
In recent years, BTEX and corresponding metabolites have been widely studied, owing to the health risks asso-ciated with exposure to these organic compounds [2 – 10]. Minor metabolites of benzene, such as trans, trans-muco-nic acid (t-MA) and S-phenylmercapturic acid (PMA), con-jugated with glutathione, are excreted in the urine. Primar-ily, alkylbenzenes are oxidized at the alkyl side chain, which in turn, give rise to aromatic carboxylic acids. Hip-puric acid (HA) and o-, m-, and p-methylhipHip-puric acid (MHA), respectively, are the chief metabolites of toluene and o-, m-, and p-xylene. Phenylglyoxylic acid (PGA) and mandelic acid (MA) are major metabolites of ethylben-zene [9]. MA, PGA, MHA, as well as a few minor
metabo-lites are not endogenously produced on a large scale. Thus, they are diagnostically more specific for determin-ing the extent of occupational or environmental exposure. The assessment of BTEX metabolite concentrations within biological materials permits the degree of exposure to these chemicals to be estimated. ACGIH (American Conference of Governmental Industrial Hygienists) pro-posed a maximum permissible level of 2.5 g of HA or 1.5 g of MHA per gram of creatinine in urine as a biological mon-itoring strategy for groups of workers [10].
Investigating BTEX metabolites requires a sufficient, rapid, sensitive, and potentially automated analytical method. To determine these metabolites in urine, high-performance liquid chromatography (HPLC) [11 – 14] and gas chromatography (GC) [15 – 16] have been applied. After the fluid samples were pretreated, quantitative ana-lyses of BTEX metabolites by HPLC or GC were per-formed. MA and PGA were analyzed by both HPLC and GC methods, both of which involved complex extraction steps and, in the case of GC, a derivatization procedure prior to the analysis [15]. The detection limit was deter-mined to be 0.020 mmol/L or less for PGA and 0.050 mol/L for MA with the HPLC method. MA and PGA in the urine of workers who are occupationally exposed to styrene were also analyzed by GC [16]. The urine sample, in an acidic medium, was extracted with chloroform and subsequently converted into the respective methyl esters. Recoveries of 66.7% and 95.4% were obtained for MA and PGA, respectively, with limits of quantitation of 0.03 g/L for MA and 0.02 g/L for PGA. Moon et al. reported
Correspondence: You-Zung Hsieh, Department of Applied Chemistry, National Chiao Tung University, Hsinchu, Tai-wan, R.O.C. Phone: +886 3 5731785. Fax: +886 3 5723764. E-mail: yzhsieh@mail.nctu.edu.tw
phase to analyze PGA, HA, and o, m, and p-MHA. The detection limits ranged from 0.6 to 2.1 mg/L with recov-eries of over 99.8% [12]. The primary disadvantages of these methods are the required extraction step, the longer analysis time, and the additional derivatization procedure in the case of GC.
Since 1990, capillary electrophoresis (CE) has proven to be of considerable value in various fields [17 – 21]. The advantages of CE are high resolution, high separation effi-ciency, lower sample consumption, and rapid analysis. Capillary zone electrophoresis (CZE) [22 – 25] and micel-lar electrokinetic chromatography (MEKC) [26 – 27] have been employed to investigate one or a few BTEX metabo-lites, but a thorough and comprehensive investigation has not been reported.
A b-cyclodextrin (b-CD) modified capillary electrophoresis method is presented herein for determining eight BTEX metabolites. In addition, the selected reference com-pounds were the endogenous urinary metabolites creati-nine and uric acid. The effects of pH and b-CD concentra-tion on the separaconcentra-tion of the analytes were examined. Finally, through a simple sample pretreatment, urine sam-ples were analyzed using the optimized method devel-oped here.
2 Experimental
2.1 Apparatus
The experiments were performed on a Beckman P/ACE system MDQ Capillary Electrophoresis System (Beckman Instruments, Fullerton, CA, USA) equipped with an UV-Vis diode-array detector. A personal computer controlled by P/ACE System MDQ software was used for data acqui-sition. Data analyses were also performed using this soft-ware. A 50 lm ID fused-silica capillary tube (Polymicro Technologies, Phoenix, AZ, USA) with a total length of 60.2 cm (50 cm to the detector) was used for the separa-tions. The capillary column was assembled in the cartridge format. The temperature of the capillary tube during the electrophoresis was maintained at 258C by means of a thermostating system. The electrophoretic separation was performed with an electric potential of 25 kV. Sam-ples were pressure injected at 0.034 bar (0.5 psi). The detection wavelength was set at 200 nm except for trans, trans-muconic acid, which was at 265 nm.
2.2 Chemicals
Borax, b-cyclodextrin, creatinine, uric acid, DL-mandelic acid, and hippuric acid were purchased from Sigma (St. Louis, MO, USA). Sodium hydroxide and sodium chloride
and p-Methylhippuric acids, trans,trans-muconic acid, and phenylglyoxylic acid were purchased from Aldrich (Mil-waukee, WI, USA). S-Phenylmercapturic acid was obtained from TCI (Tokyo, Japan). All other chemicals were of analytical grade. Water was purified with a Milli-Q water system (Millipore, Bedford, MA, USA) and filtered through a 0.22-lm filter.
2.3 Capillary electrophoresis procedures
The procedure used to condition a new capillary involved sequential treatment with 1.0 M NaOH, H2O, 0.1 M NaOH, and H2O for 30 min each. The capillary was washed between runs with 0.1 N NaOH and then water (3 min of high pressure rinsing at 138 kPa), followed by recondition-ing with the runnrecondition-ing buffer, typically for 3 min. The pH 7.2 and pH 8.1 borate running buffers were prepared by mix-ing appropriate amounts of 0.2 M boric acid and 0.1 M borax. 0.1 M NaOH was used to adjust the solutions to pH 9.0, pH 10.0, and pH 10.5 borate buffers, which con-tained appropriate amounts of 0.1 M borax.
2.4 Preparation of standard solutions and urine samples
Stock solutions of the four analytes (p-methylhippuric acid, uric acid, S-phenylmercapturic acid and trans,trans-muconic acid) were prepared in a 50 mM NaOH solution at a concentration of 3 mg/mL. The other six standards were prepared in a 50 mM NaCl solution at the same con-centration. Working standards over the range 1 to 300 lg/mL with the exception of t-MA, which was 1 to 75 lg/mL, were prepared by dilution of a standard stock solution with a 50-mM NaCl solution. Urine samples were obtained from a gas station worker who had worked there continuously for over one year and a non-exposed gradu-ate student. The urine samples were collected in the after-noon of three consecutive days. These samples were cen-trifuged at 4000 g for 10 min at room temperature, diluted 6-fold with 50 mM NaCl solution, and degassed in an ultra-sonic bath for 1 min before CE analysis. All stock and sample solutions were stored at – 208C prior to use. The recovery study followed the same procedure as described above by spiking 50 lg/mL of p-MHA, m-MHA, HA, and PGA, and the others were 10 lg/mL in gas station work-er’s urine samples.
3 Results and discussion
Figure 1 illustrates the molecular structures of eight BTEX metabolites as well as two reference compounds, creatinine and uric acid. This figure also shows that each
metabolite possesses one or two carboxyl acid groups that would be negatively charged in a basic buffer solu-tion. Moreover, there are three MHA isomers that differ relative to the positions of the substituted methyl group on the benzene ring. Except for t-MA, which absorbed at 265 nm, the maximum absorbance of the analytes was approximately 200 nm. Consequently, to determine opti-mal separation conditions, the detection wavelengths were set at 200 nm and 265 nm.
3.1 Effects of buffer pH value and b-CD concentration on the separation
Figure 2 summarizes the effects of buffer pH level on the migration behavior of the analytes, which ranged from 7.2 to 10.5. The experimental results indicate that these ana-lytes, except for creatinine, carry a negative charge within the pH range used. Notably, t-MA displayed the lowest mobility among the analytes. t-MA carried the highest negative charge density, since it possessed two carboxyl acidic groups with a relatively small molecular weight. Uric acid contains no dissociated functional groups. However, owing to its molecular structure and hydrophilic nature, it does have a negative charge to mass ratio, which is slightly less than that of t-MA. The experimental findings
revealed that with increasing pH of the running buffer, the effective mobility of MA and PGA decreased, which can be attributed to increasing proton dissociation. MA and PGA were not resolved at a buffer pH value of less than 8.1; however, they were gradually separated by increas-ing the pH of the buffer. The migration order of PGA, HA, MHA, and PMA was consistent with their increasing mole-cular size, while bearing the same charge. Notably, three MHA isomers migrated with the same mobility in the pH range employed, but failed to separate. Over a broad pH range, creatinine remained uncharged and coeluted with EOF. For the purpose of high separation efficiency, a pH 10.0 buffer was selected for further investigation. Since the three isomers of MHA failed to separate in the prescribed buffer, further improvements were required to separate all analytes.
Distinct CDs are effective modifiers, which separate the molecular structure of isomers within CE. According to the molecular structure including size, hydrophobic property, and relative position of substituents of an analyte, stable inclusion complexes can be formed with CDs. Figure 3 presents the effects of b-CD concentration on the migration behaviors of the analytes. Altering the b-CD concentration in the buffer influenced the effective mobility of most of the analytes. Furthermore, PMA was affected to the greatest degree. When b-CD was added to the running buffer, the three MHA isomers were resolved. In addition, an increased CD concentration enhanced the resolution. b-CD influenced the effective mobility of p-MHA to a greater
Figure 1. Molecular structures of the ten analytes.
Figure 2. Effect of buffer pH on the mobility of BTEX meta-bolites. 1 = creatinine; 2 = PMA; 3 = p-MHA; 4 = m-MHA; 5 = o-MHA; 6 = HA; 7 = PGA; 8 = MA; 9 = uric acid; 10 = t-MA. Conditions: capillary, 60.2 cm (50 cm to detector)650 lm ID; applied voltage, 25 kV; detection wavelength, 200 nm (before 6.5 min) and 265 nm (after 6.5 min); pressure injec-tion (3.45 kPa), 3 s; column temperature, 258C.
extent than m- and o-MHA. This suggests that p-MHA and b-CD form a relatively stable inclusion complex. It is note-worthy that an increasing b-CD concentration increased the resolution between PGA and MA. The planar structure of the substituent on PGA undergoes greater interactions with b-CD than MA. Owing to its relatively linear structure and hydrophilic property, t-MA was largely unaffected. Moreover, within the b-CD concentration range used, the migration sequence of the analytes remained unchanged. The addition of 7 mM b-CD to the pH 10.0 borax/NaOH buf-fer yielded the optimized approach to separate the ten ana-lytes, according to the peak resolution calculated for differ-ent b-CD concdiffer-entrations.
Figure 4 depicts the electropherogram of the ten analytes under the optimized separation conditions. By using 20 mM, pH 10.0 borax-NaOH buffer containing 7 mM b-CD, the eight BTEX metabolites and two reference com-pounds could be adequately separated within 10 min. The migration order of PMA, MHA, HP, and PGA adhered to the charge-to-mass ratio sequence. Since the maximum absorbance wavelength of t-MA occurred at 265 nm, the detection wavelength was altered from 200 nm to 265 nm after 6.5 min.
Figure 3. Effect of b-CD concentration on the mobility of BTEX metabolites. Conditions: separation solution, 20 mM borax-NaOH buffer, pH 10.0. Other conditions are the same as in Figure 2.
Figure 4. Separation of metabolites of BTEX under
opti-mized conditions. 1 = creatinine; 2 = PMA; 3 = p-MHA;
4 = m-MHA; 5 = o-MHA; 6 = HA; 7 = PGA; 8 = MA; 9 = uric acid; 10 t-MA. Conditions: analyte concentration, 15 lg/mL; separation solution, 20 mM borate-NaOH buffer containing 7 mM b-CD (pH 10.0); other conditions are the same as in Figure 2.
Table 1. Average migration times, repeatabilities, slopes, intercepts, correlation coefficients, and theoretical detection limits of the ten analytes.
Analyte Migration
timea)
R.S.D.a) Slopeb) Interceptb) Correlation
coefficientb) Detection limit (min) (%) (6104) (6104) (r) (lg/mL) Creatinine 3.20 0.38 0.0156 0.0380 0.999 0.20 S-Phenylmercapturic acid 4.09 0.60 0.0120 0.0220 0.999 0.30 4-Methylhippuric acid 4.45 0.66 0.0279 0.0416 0.999 0.11 3-Methylhippuric acid 4.63 0.71 0.0332 – 0.0118 0.999 0.10 2-Methylhippuric acid 4.74 0.73 0.0272 – 0.0118 0.999 0.12 Hippuric acid 4.85 0.71 0.0286 – 0.0136 0.999 0.13 Phenylglyoxylic acid 4.98 0.75 0.0228 0.0938 0.998 0.13 Mandelic acid 5.23 0.80 0.0151 0.0608 0.998 0.25 Uric acid 5.83 0.42 0.0237 0.1196 0.998 0.22 t,t-Muconic acid 9.46 0.93 0.0495 0.0705 0.999 0.10 a) n = 5, analyte concentration: 15 lg/mL.
Table 1 lists the average migration times, repeatabilities (RSDs), slopes, intercepts, linearities of the calibration graphs, as well as theoretical detection limits of the ten analytes. The migration times RSDs (n = 5) were lower than 0.93%. The metabolite concentrations were quanti-fied from the peak area of the electropherogram. Good lin-earities of calibration graphs were observed with a mini-mum of two orders of magnitude from 1 to 300 lg/mL, with the exception of t-MA, which was 1 to 75 lg/mL. The cor-relation coefficients all exceeded 0.998. At a signal-to-noise ratio of 3, the theoretical detection limits for the ana-lytes ranged from 0.10 to 0.30 lg/mL, which were lower than the data obtained previously using HPLC and GC methods [12, 15 – 16].
3.2 Determination of BTEX metabolites in urine sample
The optimized CE method was employed to determine the urinary BTEX metabolite concentrations for occupational exposure to these organic compounds. Figure 5 shows the electropherograms of urine samples from a gas sta-tion worker and a graduate student. More than ten peaks were detected in these urine samples. Four BTEX meta-bolites including p-MHA, m-MHA, HA, and PGA were detected in the urine sample from gas station worker. The migration times of four detected BTEX metabolites were 4.56, 4.73, 5.01, and 5.14 min respectively. Despite the complicated urinary matrix, those metabolites were ade-quately resolved from other unknown compounds. By comparing each peak migration time and the correspond-ing UV spectrum with those of the standard allowed those peaks to be identified. The peak purity data from the photodiode array detector of these peaks exceeded 91.3%. Moreover, to further confirm the identities of the
analytes, a standard compound, spiked into the urine sample, was employed.
MHA, m-MHA, HA, and PGA are metabolites of p-xylene, m-p-xylene, toluene, and ethylbenzene, respec-tively, which are the most volatile components of unleaded gasoline. The experimental findings indicate that, due to the occupational exposure to unleaded gaso-line in the workplace, the worker was exposed to BTEX compounds. Table 2 summarizes the BTEX metabolite content in the gas station worker’s urine sample, as well as its average recovery and relative standard deviation. The urine samples were collected for three consecutive days at approximately the same time each day. The mean
Table 2. Content and recovery of eight metabolites of BTEX in the urine sample of a gas station worker.
Analytes Day 1 Day 2 Day 3 Similarity Recovery R.S.D.
(lg/mL) (lg/mL) (lg/mL) index (%)a) (%)b) (%)b) Creatinine 1790 1770 1790 90.0 105 8.89 S-Phenylmercapturic acid –c) 97.3 3.63 4-Methylhippuric acid 49.4 56.9 61.3 91.3 90.7 2.90 3-Methylhippuric acid 41.8 32.5 29.6 99.8 90.7 2.25 2-Methylhippuric acid –c) 100 3.85 Hippuric acid 151 144 146 97.6 97.1 6.07 Phenylglyoxylic acid 24.9 29.1 36.3 99.1 93.0 1.11 Mandelic acid –c) 98.7 2.34 Uric acid 357 350 372 90.3 95.0 2.83 t,t-Muconic acid –c) 92.5 0.94
a) Peak purity data from photodiode array detector.
b) n = 3, spiked concentrations: 50 lg/mL of p-MHA, m-MHA, HA, and PGA, and the others were 10 lg/mL.
c) Not detected.
Figure 5. Electropherograms of (a) a urine sample from a gas station worker and (b) a graduate student urine sample. Other conditions and peak identities are the same as in Fig-ure 4.
90.7%. Their RSDs were less than 6.07%. This result also confirms that the b-CD modified CE method is adequate for analyzing these metabolites in urine samples. As the data in Table 2 show, none of the urinary metabolites exceeded the values recommended by ACGIH.
4 Concluding remarks
Eight BTEX metabolites were separated by applying the b-CD modified CE method with a total analysis time of 10 minutes. The advantages of analyzing BTEX metabo-lites by this method are high resolution, high separation efficiency, short analysis time, and adequate reproducibil-ity. Also, the sample needs no complicated extraction step or additional derivatization procedure. In addition, the method effectively determined the BTEX metabolites in the urine sample of a gas station worker. Therefore, the b-CD modified CE technique potentially offers a simple and rapid method for monitoring routine urine samples of per-sons who are occupationally exposed to such volatile organic compounds.
Acknowledgment
This research was supported by Grant NSC 89-2113-M-009-029 from the National Science Council of the Repub-lic of China.
References
[1] E. Reese, R.D. Kimbrough, Environ. Health Persp. 1993, 101, 115 – 131.
[2] K. Elke, E. Jermann, J. Begerow, L. Dunemann, J. Chro-matogr. A. 1998,826, 191 – 200.
[3] Y. Yang, A.D. Jones, C.D. Eaton, Anal. Chem. 1999, 71, 3808 – 3813.
[4] R. Andreoli, P. Manini, E. Bergamaschi, A. Brustolin, A. Mutti, Chromatographia 1999, 50, 167 – 172.
[5] B. Tolnai, A. Gelencsr, C. Gl, J. Hlavay, Anal. Chim. Acta 2000, 408, 117 – 122.
[6] A.-L. Nguyen, J.H.T. Luong, Electrophoresis 1997, 18, 247 – 252.
229 – 234.
[8] A. Astier, J. Chromatogr. 1993, 643, 389 – 398.
[9] J. Angerer, B. Hrsch, J. Chromatogr. 1992, 580, 229 – 255.
[10] American Conference of Governmental Industrial Hygie-nists (ACGIH), Documentation of the threshold limit values and biological exposure indices, 1995.
[11] I.F. Mao, M.L. Chen, E.W. Lo, Intern. J. Environ. Anal. Chem. 1996, 64, 1 – 9.
[12] D.H. Moon, N.W. Paik, Y.-B. Shim, J. Chromatogr. B 1997, 694, 367 – 374.
[13] A. Astier, J. Chromatogr. 1992, 573, 318 – 322.
[14] T. Fujii, S. Kawabe, T. Horike, T. Taguchi, M. Ogata, J. Chromatogr. B 1999, 730, 41 – 47.
[15] H. Kivisto, K. Pekari, A. Aitio, Int. Arch. Occup. Environ. Health 1993, 64, 399 – 403.
[16] V.L. Lanchote, A.C. Santos, R.H. Queiroz, E. Queiroz, D. Carvalho, J. Anal. Toxicol. 1994, 18, 143 – 146. [17] F. Foret, L. Krivnkov, P. Bocek, Capillary Zone
Elec-trophoresis. VCH, Weinheim 1993.
[18] N.A. Guzaman (Editor), Capillary Electrophoresis Tech-nology, Marcel Dekker, New York 1993.
[19] H. Nishi, S. Terabe, J. Chromatogr. A 1996, 735, 3 – 27. [20] H.-Y. Huang, K.-L. Kuo, Y.-Z. Hsieh, J. Chromatogr. A
1997, 771, 267 – 274.
[21] K. Peck, S.-L. Wung, G.-S. Chang, J.J.Y. Yen, Y.-Z. Hsieh, Anal. Chem. 1997, 69, 1380 – 1384.
[22] P.A. Tienstra, J.A.M. van Riel, M.D. Mingorance, C. Olie-man, J. Chromatogr. 1992, 608, 357 – 361.
[23] M. Shirao, R. Furuta, S. Suzuki, H. Nakazawa, S. Fujita, T. Maruyama, J. Chromatogr. A 1994, 680, 247 – 251. [24] C.J. Petucci, H.L. Kantes, T.G. Strein, H. Veening, J.
Chromatogr. B 1995, 668, 241 – 251.
[25] P. Simon, T. Nicot, J. Chromatogr. B 1996, 679, 103 – 112.
[26] K.-J. Lee, J.J. Lee, D.C. Moon, Electrophoresis 1994, 15, 98 – 102.
[27] J. Li, K.C. Waldron, J. Liq. Chromatogr. Rel. Techn. 1999, 22, 1919 – 1933.
