Alkaline Digestion and Solid Phase Extraction Method for Perfluorinated Compounds in Mussels and Oysters from South China and Japan

Transcription

1 Arch. Environ. Contam. Toxicol. 50, (2006) DOI: /s x Alkaline Digestion and Solid Phase Extraction Method for Perfluorinated Compounds in Mussels and Oysters from South China and Japan M. K. So, 1 S. Taniyasu, 2 P. K. S. Lam, 1 * G. J. Zheng, 1 J. P. Giesy, 1,3 N. Yamashita 2 1 Centre for Coastal Pollution and Conservation, Department of Biology and Chemistry, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong SAR, Peoples Republic ofchina 2 National Institute ofadvanced Industrial Science and Technology (AIST), 16-1 Onogawa, Tsukuba, Ibaraki , Japan 3 Department of Zoology, National Food Safety and Toxicology Center, Center for Integrative Toxicology, Michigan State University, East Lansing, Michigan 48824, USA Received: 10 March 2005 /Accepted: 8 May 2005 Abstract. Perfluorinated compounds (PFCs), such as perfluorooctane sulfonate (PFOS), have been identified in the coastal waters ofchina and Japan. An alkaline digestion method, coupled with solid-phase extraction (SPE), and highperformance liquid chromatography interfaced with high-resolution electrospray tandem mass spectrometry was developed to determine PFCs in mussel and oyster samples from coastal waters ofsouth China and Japan. These techniques produced adequate recoveries and reporting limits with small quantities ofpfcs. Concentrations ofindividual PFCs in mussels and oysters from south China and Japan ranged from to pg/g, wet weight (ww) for PFOS, 63.1 to pg/g, ww for perfluorohexane sulfonate, 9.3 to 30.1 pg/g, ww for perfluorobutane sulfonate and 37.8 to pg/g, ww for perfluorooctane sulfonamide. The quantification of perfluorinated carboxylates was compromised by interferences from carboxylates in the procedural blanks. Perfluoroundecanoate and perfluorononanoate had relatively great blank interferences, which resulted in relatively poor limits of quantification for these compounds. Some PFCs were only identified in a limited number of samples: perfluorododecanoate in samples from Tokyo Bay, Japan (195.9 pg/g, ww); and perfluorodecanoate in Fuzhou, China (131.7 pg/g, ww) and Tokyo Bay (118.6 pg/g, ww). The greatest concentrations ofperfluorooctanoate, perfluoroheptanoate, and perfluorohexanoate were observed in samples from Tokyo Bay and Bei Hai, south China. The extensive application ofperfluorinated compounds (PFCs) in many commercial products has led to the widespread occurrence ofpfcs in the environment, and PFCs are found globally in tissues ofmany wildlife species, and in blood serum ofboth occupationally and non-occupationally exposed people (Hoff 2003; Kannan et al. 2001; Kuklenyik et al. 2004; *Correspondence to: P. K. S. Lam; bhpksl@cityu.edu.hk Olsen et al. 1999, 2003; Van de Vijver et al. 2003, 2004). The widespread occurrence and potential toxicities ofsome PFCs, have led to their production being voluntarily phased out in 2001 (Martin et al. 2004a). Among the PFCs, perfluorooctane sulfonate (PFOS), perfluorooctanoate (PFOA), perfluorooctane sulfonamide (PFO- SA), perfluorohexane sulfonate (PFHS), and perfluorobutane sulfonate (PFBS) have received the most attention (Tomy et al. 2004). PFOS has been found in almost every wildlife species in which it has been measured. Although PFOS and PFOA have been measured most frequently, long-chained perfluorocarboxylates (PFCAs), with carbon numbers ranging from 9 to 15, have been found in ng/g, wet weight (ww) concentrations in liver samples ofvarious species from the Canadian Arctic. In mink liver, the mean perfluorononanoate (PFNA) (16 ng/g, ww) concentration was greater than that ofpfos (8.7 ng/g, ww) (Martin et al. 2004b). Longer-chained PFCAs [the series from perfluorohexanoate (PFHxA) to 14-carbon version of PFOA (perfluorotetradecanoate, PFTA)] have also been observed in the fish liver tissues collected from Etobicoke Creek, Toronto, Canada, where an accidental release of fire-fighting foam occurred (Moody et al. 2002). The possibility ofbioaccumulation and bioconcentration oflong-chained perfluoroalkyl carboxylates and sulfonates (Martin et al. 2003a, 2003b) emphasizes the necessity ofmonitoring and evaluating the toxicological effects of these compounds in the marine environment. In order to monitor these PFCs, a reliable method with acceptable precision and accuracy is needed. Many previous studies utilized ion-pairing methods for extracting PFCs in blood, liver, and tissue samples. Those methods, however, gave poor recoveries for some of the analytes (Martin et al. 2004a). A new automated solid-phase extraction and measurement method for PFCs in serum and milk achieved acceptable recoveries for most analytes, except perfluoropentanoate (PFPeA) and perfluorododecanoate (PFDoDA), of which the recoveries were poor (22% and 33% in serum, respectively) (Kuklenyik et al. 2004). Most ofthe published sample preparation methods do not involve any special

2 241 Perfluorinated Compounds in Mussels, Oysters from China and Japan cleanup procedures, which results in interferences during instrumental analysis (Martin et al. 2004a). Solid-phase extraction (SPE) shows promise as a reliable and relatively simple cleanup technique, especially for biological samples (Kuklenyik et al. 2004; Martin et al. 2004a). The aim of this study was to develop a practical method for measuring a broad range of PFCs in biological matrices. With rapid industrialization and development in the past decade, the coastal waters of south China and Hong Kong have been increasingly influenced by pollutant exposures from a wide range of industrial and domestic sources (So et al. 2004). Previous studies of these coastal waters have primarily focused on traditional persistent organic pollutants including PCBs, DDTs, PAHs, and other OC pesticides (Fu et al. 2003; Mai et al. 2002; Yang et al. 1997). A recent study revealed that concentrations of PFCs, mainly PFOS and PFOA, in the coastal waters of south China and Hong Kong were similar to those found in the coastal marine environment of Japan and south Korea (So et al. 2004). Ranges of concentrations of PFOS in Hong Kong, and the Pearl River Delta, China, were and ng/l, respectively, whereas those for PFOA were and ng/l, respectively (So et al. 2004). Although all the concentrations were less than those that would be expected to cause adverse effects to aquatic organisms or their predators, the presence of PFCs in coastal waters suggests their widespread distribution and potential releases from adjacent industrialized areas in Hong Kong and south China. Further monitoring and identification of sources of PFCs are necessary for a more comprehensive risk assessment. Fig. 1. The sampling locations along the East Coast of China and in Japan: Qinzhou (QZ), Fang Cheng (FC), Bei Hai (BH), Xiamen (XI), Fuzhou (FZ), SengSi Dao (SS), Tokyo Bay (TB) methanol before use. Disposable PP copolymer laboratory ware was used to minimize potential analyte loss. Teflon and glass containers were avoided; the former cause analytical interferences, and the latter may bind the surfactants in an aqueous solution. Sample Preparation Materials and Methods Chemicals and Standards External standards, a mixture of perfluorinated compounds (PFCmix) containing PFOS, PFHS, PFBS, PFOSA, PFDoDA, perfluoroundecanoate (PFUnDA), perfluorodecanoate (PFDA), PFNA, PFOA, perfluoroheptanoate (PFHpA) and PFHxA, at 100 ng/ml were supplied by the National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Japan. Methanol (high-performance liquid chromatography (HPLC) grade), distilled water (HPLC grade), potassium hydroxide (KOH, 85%), and ammonium acetate were purchased from Wako Pure Chemical Industries, Ltd (Osaka, Japan). Nylon mesh filters (12 lm, 13 mm i.d.) were purchased from Iwaki (Tokyo, Japan). All solvents and reagents were used as received. Frozen tissue samples from each location were thawed and homogenized by a mechanical homogenizer in a PP bottle. After homogenizing duplicate samples from one location, the homogenizer probe and laboratory ware were thoroughly washed in the sequence of tap water, distilled water, and methanol before homogenizing the next set of samples to avoid cross contamination. Approximately 1 g of homogenized tissue was transferred to a 50-mL PP centrifuge tube, and 150 ll of 100 ng/ml PFCmix and 30 ml of 0.01 N KOH/ methanol were added to the PP tube. The mixture was shaken at 250 rpm at room temperature ( 25 oc) for 16 h. After digestion, 1 ml of the tissue solution mixture was added to a 1-L PP bottle containing 100-mL distilled water and was shaken thoroughly. This tissue water mixture was then extracted for subsequent analysis. Sample Extraction by Solid Phase Extraction Cartridge Sample Collection Green-lipped mussels, Perna viridis, were collected from six locations, namely, Qinzhou (QZ), Fang Cheng (FC), Bei Hai (BH), Xiamen (XI), Fuzhou (FZ) and Seng Si Dao (SS) along the east coast of China in April Oysters (Crassostrea gigas) were also collected from Tokyo Bay (TB) in Japan in July 2004 (Figure 1). All the samples were stored in polyethylene bags, kept in ice, and transported to the laboratory immediately. Mussel and oyster soft tissues were removed from shell and placed in polypropylene (PP) tubes. The samples were stored at 20 C until analyzed. All equipment, storage containers, and laboratory ware were cleaned by rinsing with distilled water followed by The tissue water homogenate was extracted by using an Oasis HLB cartridge (Oasis HLB: 0.2 g, 6 cm3; Waters Corp. Milford, Massachusetts). The procedures are a slight modification of a previously described method (So et al. 2004). Each sample was extracted with a separate cartridge. Prior to loading the cartridge with a sample, the cartridge was preconditioned by eluting with 5 ml of methanol followed by 5 ml of distilled water at a rate of 2 drops/sec. The tissue water mixture was then loaded onto the cartridge, eluting at a rate of 1 drop/sec, and the eluate was discarded. The cartridge was prevented from becoming dry at all times during the loading of the samples. As part of method development, a separate and preliminary washing step was evaluated: each HLB cartridge was washed with 5

3 242 M. K. So et al. Table 1. Multiple-reaction-monitoring analysis ofpfcs indicating precursor and product negative ion masses and MSMS conditions Analyte Abbreviation Precursor ion Product ion Collision energies (ev) Cone voltages (V) Perfluorooctanesulfonate PFOS Perfluorobutanesulfonate PFBS Perfluorohexanesulfonate PFHS Perfluorooctanesulfonamide PFOSA Perfluorohexanoate PFHxA Perfluoroheptanoate PFHpA Perfluorooctanoate PFOA Perfluorononanoate PFNA Perfluorodecanoate PFDA Perfluoroundecanoate PFUnDA Perfluorododecanoate PFDoDA ml of40% methanol in water at a rate of1 drop/sec, and the eluate was collected in a 15-mL PP centrifuge tube (Fraction 1). Following this step, the cartridge was allowed to run dry and the remaining water was completely removed from the cartridge by using a PP syringe. A second fraction (Fraction 2) was eluted with 10 ml of 100% methanol at a rate of1 drop/sec. The eluates in Fraction 1 and Fraction 2 were then reduced to 3 ml and 1 ml, respectively, under a gentle stream of high-purity nitrogen gas. After volume reduction, the eluates were passed through separate 0.1 lm 13 mm nylon mesh filters (which were pre-washed with ml methanol), and then transferred into the corresponding 15-mL PP centrifuge tube. Any eluates that remained in the nylon filter were washed through the filter with 0.5 ml of100% methanol. The final volume of the eluate was reduced by high-purity nitrogen gas to 3 ml for Fraction 1 and 0.5 ml for Fraction 2. One hundred microliters ofeluate was transferred to a PP vial for instrumental analysis. Instrumental Analysis Concentrations ofpfcs in tissue samples were analyzed using HPLC with high-resolution, electrospray tandem mass spectrometry (HPLC- MS/MS). Separation ofanalytes was performed by an Agilent HP1100 liquid chromatograph (Agilent, Palo Alto, California) interfaced with a Micromass Quattro II mass spectrometer (Waters Corp., Milford, Massachusetts) operated in electrospray negative mode. A 10-lL aliquot ofextract was injected onto a Keystone Betasil C 18 column (2.1 mm i.d. 50-mm length, 5 lm) with 2 mm ammonium acetate and methanol as the mobile phases starting at 10% methanol at a flow rate of300 ll/min. The gradient was increased to 100% methanol at 10 min before reverting to the original conditions at 12 min. The desolvation gas flow and temperature were kept at 750 L/h and 400 C. The collision energies, cone voltages, and MS/MS parameters for the instrument were optimized individually for each analyte and are summarized (Table 1). Data Analysis Acquisition analysis for all of the samples, blanks, and QC materials were performed using MassLynx software of the Micromass Quattro II system (Waters Corp., Milford, MA). Each ion of interest in the chromatogram was automatically selected and integrated. Concentrations of target analytes were quantified by using calibration curves constructed using external standards of six different concentrations (2, 10, 50, 200, 1000, and 20,000 ng/l). Apart from a few outliers, the standard calibration curve showed strong linearity (correlation coefficients >0.99). Prior to sample quantification, the instrument was tuned to obtain the maximum peak resolution and the greatest mass accuracy. Multiple reaction monitoring (MRM) mode, having a higher sensitivity, was used in the present study for quantification. In MRM mode, a daughter ion, usually the one with the greatest intensity, produced from a selected parent ion was monitored and quantified. Results and Discussion Optimization of Alkaline Digestion Due ofthe specific protein-binding properties ofpfcs, biological matrix interference reduces the sensitivity of the analytical technique. Therefore, it is necessary to remove the proteins to release PFCs for increased sensitivity. Alkaline digestion oflipids and proteins before extraction was suggested to achieve an accurate and reliable measurement of PFCs in biological samples. Current sample preparation methods do not include a cleanup step to remove these matrix interferences (Martin et al. 2004a). To evaluate the effect of alkaline digestion on target analytes, 10 ll of100 ng/ml PFCmix was added to 1 ml of different concentrations ofkoh in water (0.01, 0.1, 0.3, 0.5, 1, and 2 N) or methanol (0.01, 0.1, and 0.3 N). The percentage recoveries for all the target analytes under different KOH concentrations are summarized in Table 2. In general, all alkaline/ water mixtures gave poor recoveries for PFDoDA and PFUnA and other PFCs, and were judged unacceptable for the alkaline digestion protocol. The alkaline/methanol solution (0.01 N KOH/methanol) was found to have the most consistent recoveries for all PFCs (>70%). For many PFCs, greater KOH concentrations reduced recovery rates. The duration for the alkaline digestion process was also tested by shaking approximately 1 g ofhomogenized mussel tissue sample with 30 ml 0.01 N KOH/methanol solution. Small particles were still observed after shaking for 2 and 6 h. Consequently, the samples were digested overnight (10 h) to ensure complete digestion ofthe fat and proteins in the homogenized tissue. Quality Assurance and Control BlankAnalysis The occurrence ofrelatively great concentrations ofsome PFCs in blanks is a persistent problem in the analyses ofthese

4 Perfluorinated Compounds in Mussels, Oysters from China and Japan 243 Table 2. The effect ofconcentration ofkoh/methanol on the recoveries ofpfcs Concentration (ng/l) Conc. OfKOH in methanol (KOH/M) or water (KOH/W) (N) PFOS PFHS PFBS PFOSA PFDoDA PFUnDA PFDA PFNA PFOA PFHpA PFHxA 0.01NKOH/W NKOH/W NKOH/W NKOH/W NKOH/W NKOH/W NKOH/M NKOH/M NKOH/M Table 3. Background level of perfluorinated compounds (pg) in HPLC grade methanol, nylon mesh filter, and the whole procedures Concentration (pg) PFOS PFHS PFBS PFOSA PFDoDA PFUnDA PFDA PFNA PFOA PFHpA PFHxA <1.6 <6.7 <1.6 <1.6 <1.6 <1.6 <1.6 <1.6 <1.6 <1.6 <1.6 Nylon filter a Without washing 27.6 <2 <2 3.3 <2 <2 < <50 <10 With washing <10 <2 2.7 <2 <2 <2 < <10 <50 <10 Procedural blank Fraction 1 b < <8.2 <43.6 <11.8 < < Fraction 2 c <10 <2.6 <2.6 < <36.8 <50 <31.3 a Amount in the final 1 ml solution and 10 ll was injected into the LC/MS/MS for quantification. b Amount in the final 3 ml solution and 10 ll was injected into the LC/MS/MS for quantification. c Amount in the final 0.5 ml solution and 10 ll was injected into the LC/MS/MS for quantification. compounds. The sources ofcontamination can be laboratory products containing fluoropolymers such as polytetrafluoroethylene (Martin et al. 2004a). The difficulties presented by background contamination ofpfcs in blank test matrix was evaluated and minimized by the following procedures: (1) elimination ofteflon-containing equipment, (2) use ofprecleaned and methanol-rinsed glassware, and (3) use ofpp storage vials. Contamination ofequipment and solvent were tested for presence of background concentrations oftarget analytes. The methanol, nylon mesh filter, as well as the entire procedure, were evaluated as sources ofbackground levels of target analytes before samples were analyzed. Evaluation of methanol contamination was performed by concentrating 30 ml ofmethanol to 1 ml by gentle nitrogen blown-down and then injected into the LC/MS/MS for analysis. Concentrations for all of the compounds of interest were less than the limit of quantification (LOQ), indicating that HPLC-grade methanol would not contribute much interference for PFC analysis (Table 3). Contamination from the nylon mesh filter was evaluated by comparing prewashed versus nonwashed filters using standard cleaning procedures. Pre- and nonwashed filters were rinsed with 1 ml of100% methanol; the eluate was collected and analyzed. Trace amounts ofpfos (27.6 pg) and PFOA (34.6 pg) were found in the nontreated nylon filter. The levels of these compounds in the background were greatly reduced by washing the filter with methanol prior to sample loading (Table 3). In previous analyses ofwater (So et al. 2004), direct SPE extraction ofpfcs without digestion was utilized to evaluate coastal waters ofhong Kong, south China, and Korea. However, in the presence ofa biological matrix, the technique had to be modified with alkaline digestion to attempt to reduce matrix interferences. Procedural blank, following the same extraction procedure, was carried out for each set of samples to test for possible contamination. Considerable amounts of PFDoDA (8.2 pg), PFUnDA (43.6 pg), PFDA (11.8 pg), and PFNA (61.8 pg) were detected in the target fraction (F2), while concentrations ofall the other chemicals were less than the LOQ. Trace amounts ofpfoa (36.8 pg) and PFHxA (31.3 pg) were also found in Fraction 1 (Table 3). The LOQ for each target analyte was determined as the smallest mass ofstandard injected that resulted in a reproducible measurement ofpeak areas consistent with the calibration curve. The LOQs (amount in the final 0.5 ml solution and 10 ll was injected for quantification) were 10 pg for PFOS, 2.6 pg for PFHS and PFBS, 5.1 pg for PFOSA, 8.2 pg for PFDoDA, 43.6 pg for PFUnDA, 11.8 pg for PFDA, 61.8 pg for PFNA, 36.8 pg for PFOA, 50 pg for PFHpA, and 31.3 pg for PFHxA. Spike Recovery Analysis Recoveries ofpfcs were performed to determine the precision and accuracy ofthe extraction and analytical procedures. Two

5 244 M. K. So et al. Table 4. Recoveries (%) for individual PFCs with and without alkaline digestion Recovery (%) Without alkaline digestion (n = 2) With alkaline digestion (n = 2) Fraction 1 Fraction 2 Fraction 1 Fraction 2 Fraction 1+2 PFOS PFHS PFBS PFOSA PFDoDA PFUnDA PFDA PFNA PFOA PFHpA PFHxA Note: The values are mean standard deviations. separate spike recovery evaluations are reported, which evaluate (1) the alkaline digestion effect and (2) the biological matrix effect. Alkaline Digestion. To determine the recoveries ofpfcs with alkaline digestion, 150 ll of100 ng/ml PFCmix was spiked into 30 ml KOH/methanol, digested for 16 h, and then extracted following the same procedures used for the biological samples. The recoveries ofpfcs without alkaline digestion were studied following a similar procedure, except that the 150 ll PFCmix was added directly to 100 ml distilled water. The recoveries for all the target analytes are summarized (Table 4). Without alkaline digestion, most of the chemicals were detected in Fraction 2 with recoveries of greater than 70%. The recoveries in Fraction 1, usually regarded as a washing step, were less than 0.5% for most of the compounds. Recoveries ofvarious PFCs after alkaline digestion are given (Table 4). The recoveries in Fraction 2 were 149.1% for PFOS, 33.2% for PFHS, 104% for PFOSA, 91.6% for PFDoDA, 110.5% for PFUnDA, 114.7% for PFDA, and 101.1% for PFNA. Relatively great recoveries were also detected in Fraction 1 for perfluorinated sulfonic acids (PFHS and PFBS) and carboxylic acids (PFOA, PFHpA and PFHxA) with carbon chain lengths ofless than 6 and 8, respectively. The detection oftarget analytes in Fraction 1 may be attributed to (1) alterations in the structure ofthese compounds after alkaline digestion; and/or (2) the effect of KOH/ methanol on the adhesion capacity ofthe analytes to the cartridge sorbent. By combining Fractions 1 and 2, most of the compounds achieved recoveries ofmore than 92%. On the basis ofthe above observations, the washing step (Fraction 1) was eliminated from final analytical scheme, and all target PFCs were eluted directly with 15 ml of100% methanol. Matrix Effects. The effect of biological matrix on the recoveries ofpfcs was investigated with a matrix spike recovery test. Approximately 1 g ofhomogenized oyster tissue collected from Tokyo Bay was spiked with 150 ll of 100 ng/ml PFCmix and digested by 30 ml KOH/methanol for 16 h, and then extracted following the same procedures as described above (n = 2). A second series ofduplicate samples, without the addition ofoyster homogenate, was processed in the same manner. The recoveries ofvarious PFCs from this spike exercise are presented (Table 5). In general, recoveries ofpfcs were comparable between samples with or without oyster tissue present; recoveries ofall PFCs exceeded 76% and were judged acceptable as quantitative estimates ofpfc concentrations in tissues, except for PFOA, which had a high recovery in samples without oyster tissue (197%) compared with those with oyster tissue (93%). This high recovery was probably the result ofpfoa contamination during the extraction, and further analysis is required to ascertain this point. Notwithstanding, results ofthe present study indicated that the sample matrix does not cause a significant reduction in concentrations spiked onto the oyster tissue matrix. Further study should examine the matrix effects of other tissue types.!?a3b2 tpb=3.8mm?> Linearity The accuracy ofthe above method was further assessed by performing linearity tests using four different amounts of oyster tissues (0.8, 1.1, 1.7, 2.0 g) for extraction. The correlation between increased amounts ofanalyte with proportional increases oftissue weight is evaluated by linear regression (Table 6). Good linearity was found between amounts ofoyster tissues and several PFCs including PFOS, PFBS, PFUnDA, and PFOA. The r 2 values were 0.73 for PFOS, 0.94 for PFBS, 0.70 for PFUnDA, and 0.89 for PFOA. The r 2 values for PFHS, PFDA, and PFNA ranged from 0.56 to 0.58, and those for PFOSA, PFDoDA, PFHpA, and PFHxA were less than The apparent lack of strong linear relationship for these latter compounds is probably related to the small quantities ofoyster tissues used, and the relatively small concentrations ofthese PFCs in the tissues.

6 Perfluorinated Compounds in Mussels, Oysters from China and Japan 245 Table 5. Recoveries (%) for individual PFCs with and without oyster samples Recovery (%) Without oyster samples (n = 2) With oyster samples (n = 2) Fraction 1 Fraction 2 Fraction Fraction 1 Fraction 2 Fraction PFOS PFHS PFBS PFOSA PFDoDA PFUnDA PFDA PFNA PFOA PFHpA PFHxA The values are mean standard deviations. Table 6. Linearity test between amounts oftissue weights and concentration ofpfcs Target analyte Correlation coefficient (r 2 ); n = 4 PFOS 0.73 PFHS 0.56 PFBS 0.94 PFOSA 0.43 PFDoDA 0.50 PFUnDA 0.70 PFDA 0.58 PFNA 0.56 PFOA 0.89 PFHpA 0.06 PFHxA 0.22 Modification and Finalization of Solid Phase Extraction Method As a result ofthe presence ofcertain target analytes in Fraction 1, the preliminary washing step was eliminated. The cartridge was directly eluted with 15 ml (instead of10 ml) 100% methanol after sample loading, and the final volume was reduced to 0.2 ml instead of0.5 ml. All the other procedures were not altered. With the change in solvent volume and washing process, solution spike recovery after alkaline digestion was performed to determine the accuracy and precision of the modified method. A matrix spike recovery test was not carried out at this time because the results ofthe solution spike recovery evaluation were consistent with the earlier evaluation. Procedural blanks were carried out for each set of extraction to test for possible sources of contamination. The recovery (%) and the level ofblank (pg) in 0.2 ml final extract solution are summarized (Table 7). Except for a trace amount ofperfluorinated carboxylic acid (PFDoDA, PFUnDA, and PFNA) found in the procedural blank, all the other analytes were less than the LOQ. The recoveries for the target analytes were more than 75%, except for PFDoDA, which had a recovery of57.4%. Table 7. Recoveries (%) and level ofblank (pg) for individual PFCs. Level ofblank (pg)* Recovery (%) Mean S.D. Mean S.D. PFOS < PFHS < PFBS < PFOSA < PFDoDA PFUnDA PFDA < PFNA PFOA < PFHpA < PFHxA < Amount in 0.2 ml final extract solution and 10 ll was injected into the LC/MS/MS for quantification. Most ofthe previous studies utilized an ion-pairing method for the quantitation ofa limited number ofpfcs from biological samples such as liver and serum. It was not until 2003 that a study measured the concentrations ofa much wider range of PFCs from biological tissue samples (fish carcasses) (Martin et al. 2003a). The recoveries were approximately 95% for PFHpA, 100% for PFOA, 80% for PFDA, PFUnA, PFHxS and PFOS, 65% for PFDoDA as well as 55% for PFTA and PFBS (Martin et al. 2003a). In the present study, except for PFDoDA, all the other PFCs had recoveries at least comparable to, ifnot higher than, that ofprevious study. In particular, the mean recovery for PFBS (84.7%) was much higher in the present study. In addition, the method described here also allowed the analysis ofpfhxa with a good recovery (94.5%). Application of the Technique to Mussel and Oyster Samples from China and Japan The modified method was used to analyze mussel samples along the east coast ofchina, and oyster samples collected from Tokyo Bay in Japan. Concentrations of PFCs, in terms of

7 246 M. K. So et al. Table 8. Concentrations ofpfcs (pg/g, wet weight and dry weight) in mussel and oyster samples collected from East Chinese Coast and Tokyo Bay in Japan Concentration (pg/g) Moisture Location content (%) PFOS PFHS PFBS PFOSA PFDoDA PFUnDA PFDA PFNA PFOA PFHpA PFHxA Wet weight basis FZ < <102.0 < < <204.0 <300.0 <183.0 XI < <102.0 < <114.0 < <204.0 <300.0 <183.0 QZ <102.0 < <114.0 < <204.0 <300.0 <183.0 FC <102.0 < <114.0 < <204.0 <300.0 <183.0 SS <102.0 < <114.0 < BH <102.0 < <114.0 < TB < < < Dry weight basis FZ < <796.9 < <12,468.8 < < < XI < <772.7 < <863.6 <12,090.0 < < < QZ <582.9 < <651.4 < < < < FC <772.7 < <863.6 <12,090.0 < < < SS <803.1 < <897.6 <12, BH <551.4 < <616.2 < TB < <10, < The values are mean standard deviations. Fig. 2. Composition of perfluorinated sulfonic acid and sulfonamide in mussel and oyster samples from different sampling locations pg/g, in mussel and oyster samples are summarized in Table 8. Unless otherwise stated, all concentrations ofpfcs in mussel or oyster tissues measured in this study are expressed on a wet weight basis. Most ofthe mussel and oyster samples contained detectable concentrations ofpfos, PFHS, PFBS, and PFOSA. Concentrations ranged from to pg/g for PFOS, 63.1 to pg/g for PFHS, <12.0 to 30.1 pg/g for PFBS, and 37.8 to pg/g for PFOSA. Oyster samples from Tokyo Bay had the greatest concentrations ofpfos (586.0 pg/g) and PFOSA ( pg/g). Nevertheless, the level ofpfos [ pg/g (recalculated on a dry weight basis)] was much less than the greatest PFOS concentration (1,106,000 pg/g dry weight) measured in oysters from Hog Point, Chesapeake Bay (Kannan et al. 2002). The comparatively small concentrations of perfluorinated sulfonic acid and sulfonamide in mussel samples from China indicated that the east coast of China is probably not severely polluted by PFCs. In general, concentrations of perfluorinated sulfonic acids (expressed on a dry weight basis) were in a similar range as PCBs ( ,000 pg/g), chlordanes ( pg/g), and hexachlorocyclohexanes ( pg/g), but less than that ofddts (14, ,000 pg/g), in mussels collected along the east coast of China (Monirith et al. 2003). The composition ofperfluorinated sulfonic acids and sulfonamide were compared among different sampling locations (Figure 2). Similar compositions were found in mussels from Bei Hai, Seng Si Dao, Fang Cheng, and Qingzhou, with PFOS as the major component. A completely different composition profile was observed in oysters from Tokyo Bay, where PFOSA accounted for approximately 80% ofthe PFCs. Assuming that the uptake of different PFCs by mussels and oysters followed a similar mechanism, these findings suggest that the sources of PFC contamination in Japan and China could be quite different. In Japan, the high PFOSA concentrations could probably be due to the extensive use ofinsecticide to control termites and ants. In Fuzhou, PFHS showed the greatest concentration among all the PFCs, indicating that a source ofpfhs contamination might exist in this area. The analysis ofpfcas in the present study had two major technical difficulties: (1) trace amounts of carboxylic acids, especially PFUnDA (210 pg/ml) and PFNA (266 pg/ml), were found in the procedural blank; and (2) the concentrations ofseveral PFCs in the samples were comparable to or only slightly greater than those found in the procedural blank.

8 Perfluorinated Compounds in Mussels, Oysters from China and Japan 247 Fig. 3. The comparison oftotal concentration ofperfluorinated carboxylic and sulfonic compounds in mussel and oyster samples from individual sampling location Concentrations ofpfunda and PFNA in all the samples were less than the LOQ. PFDoDA was only detected in oysters from Tokyo Bay (195.9 pg/g), whereas PFDA was detected in samples from Fuzhou (131.7 pg/g) and Tokyo Bay (118.6 pg/ g). The greatest concentrations ofpfoa were found in samples from Tokyo Bay, whereas the greatest concentrations of PFHpA and PFHxA were found in samples from Bei Hai. The total concentrations of perfluorinated carboxylic and sulfonic compounds in mussels and oysters from individual sampling locations were compared (Figure 3). Total concentrations of sulfonic compounds were higher than carboxylic compounds in samples from Fuzhou and Tokyo Bay and vice versa for those from Seng Si Dao and Bei Hai. This pattern indicates that sulfonyl-containing compounds were or are commonly used in areas near Fuzhou and Tokyo Bay. Although a bioconcentration study ofpfcs on rainbow trout demonstrated that the bioconcentration potential of PFCs increased with increasing perfluoroalkyl chain length (Martin et al. 2003a), such a pattern was not observed for perfluorinated carboxylates in the present study. Except for the Fuzhou mussels, samples from other locations showed greater concentrations oflonger-chained perfluorinated sulfonates. The reason for the absence of such a pattern for carboxylates is not known, but could be due to different organisms and different environments that were investigated in the respective studies. Summary Alkaline digestion, coupled with Oasis HLB solid-phase extraction, were developed and evaluated for a range of long-chained and short-chained PFCs from biota samples. This rapid and relatively simple extraction method performed as well as other published methods for tissue analysis ofpfcs. Acceptable recoveries were achieved for most ofthe target analytes, but trace amounts ofcertain PFCs in the procedural blank restricted the applicability ofthis method to samples containing relatively high concentrations with respect to those compounds. Further work will be required to minimize the blank level by improving the solid phase extraction procedure. Acknowledgment. This study was supported by the Area ofexcellence Scheme under the University Grants Committee ofthe Hong Kong Special Administration Region, China (Project No. AoE/P-04/ 2004), and a RGC-CERG (cityu1401/05m). References Fu J, Mai B, Sheng G, Zhang G, Wang X, Peng P, Xiao X, Ran R, Cheng F, Peng X, Wang Z, Tang UW (2003) Persistent organic pollutants in environment ofthe Pearl River Delta, China: An overview. Chemosphere 52: Hoff PT, Van de Vijver K, Van Dongen W, Esmans EL, Blust R, De Coen WM (2003) Perfluorooctane sulfonic acid in bib (Trisopterus luscus) and plaice (Pleuronectes platessa) from the Western Scheldt and the Belgian North Sea: Distribution and biochemical effects. Environ Toxicol Chem 22: Kannan K, Koistinen J, Beckmen K, Evans T, Gorzelany JF, Hansen KJ, Jones PD, Helle E, Nyman M, Giesy JP (2001) Accumulation of perfluorooctane sulfonate in marine mammals. Environ Sci Technol 35: Kannan K, Hansen KJ, Wade TL, Giesy JP (2002) Perfluorooctane sulfonate in oysters, Crassostrea virginica, from the Gulf of Mexico and the Chesapeake Bay, USA. Arch Environ Contam Toxicol 42: Kuklenyik Z, Reich JA, Tully JS, Needham LL, Calafat AM (2004) Automated solid-phase extraction and measurement ofperfluorinated organic acids and amides in human serum and milk. Environ Sci Technol 38: Mai BX, Fu JM, Sheng GY, Kang YH, Lin Z, Zhang G, Min YS, Zeng EY (2002) Chlorinated and polycyclic aromatic hydrocarbons in riverine and estuarine sediments from Pearl River Delta, China. Environ Pollut 117: Martin JW, Mabury SA, Solomon KR, Muir DCG (2003a) Dietary accumulation ofperfluorinated acids in juvenile rainbow trout (Oncorhynchus mykiss). Environ Toxicol Chem 22: Martin JW, Mabury SA, Solomon KR, Muir DCG (2003b) Bioconcentration and tissue distribution ofperfluorinated acids in rainbow trout (Oncorhynchus mykiss). Environ Toxicol Chem 22: Martin JW, Kannan K, Berger U, Voogt PD, Field J, Franklin J, Giesy JP, Harner T, Muir DCG, Scott B, Kaiser M, Järnberg U, Jones KC, Mabury SA, Schroeder H, Simcik M, Sottani C, Bavel BV, Kärrman A, Lindstrçm G, Leeuwen SV (2004a) Analytical challenges hamper perfluoroalkyl research. Environ Sci Technol 248A 255A Martin JW, Smithwick MM, Braune BM, Hoekstra PF, Muir DCG, Mabury SA (2004b) Identification of long-chain perfluorinated acids in biota from the Canadian Arctic. Environ Sci Technol 38: Monirith I, Ueno D, Takahashi S, Nakata H, Sudaryanto A, Subramanian A, Karuppiah S, Ismail A, Muchtar M, Zheng J, Richardson BJ, Prudente M, Hue ND, Tana TS, Tkalin AV, Tanabe S

9 248 M. K. So et al. (2003) Asia-Pacific mussel watch: monitoring contamination of persistent organochlorine compounds in coastal waters ofasian countries. Mar Pollut Bull 46: Moody CA, Martin JW, Kwan WC, Muir DCG, Mabury SA (2002) Monitoring perfluorinated surfactants in biota and surface water samples following an accidental release of fire-fighting foam into Etobicoke Creek. Environ Sci Technol 36: Olsen GW, Burris JM, Mandel JH, Zobel LR (1999) Serum perfluorooctane sulfonate and hepatic and lipid clinical chemistry tests in fluorochemical production employees. J Occup Environ Med 41: Olsen GW, Hansen KJ, Stevenson LA, Burris JM, Mandel JH (2003) Human donor liver and serum concentrations ofperfluorooctanesulfonate and other perfluorochemicals. Environ Sci Technol 37: So MK, Taniyasu S, Yamashita N, Giesy JP, Zheng J, Fang Z, Im SH, Lam PKS (2004) Perfluorinated compounds in coastal waters of Hong Kong, South China and Korea. Environ Sci Technol 38: Tomy GT, Tittlemier SA, Palace VP, Budakowski WR, Braekevelt E, Brinkworth L, Friesen K (2004) Biotransformation of N- ethyl perfluorooctanesulfonamide by rainbow trout (Onchorhynchus mykiss) liver microsomes. Environ Sci Technol 38: Van de Vijver KI, Hoff PT, Van Dongen W, Esmans EL, Blust R, De Coen WM (2003) Exposure patterns ofperfluorooctane sulfonate in aquatic invertebrates from the Western Scheldt estuary and the Southern North Sea. Environ Toxicol Chem 22: Van de Vijver KI, Hoff PT, Das K, Van Dongen W, Esmans EL, Siebert U, Bouquegneau JM, Blust R, De Coen WM (2004) Baseline study ofperfluorochemicals in harbour porpoises (Phocoena phocoena) from Northern Europe. Mar Pollut Bull 48: Yang YH, Sheng GY, Fu JM, Min YS (1997) Organochlorinated compounds in waters ofthe Pearl River Delta region. Environ Monit Assess 44: