Cloud-Point Extraction and Preconcentration of Cyanobacterial Toxins

A new cloud-point extraction and preconcentration method using a cationic surfactant, Aliquat-336 (tricaprylyl- methylammonium chloride), has been dev...
0 downloads 7 Views 121KB Size
Download PDF
Environ. Sci. Technol. 2002, 36, 3985-3990

Cloud-Point Extraction and Preconcentration of Cyanobacterial Toxins (Microcystins) from Natural Waters Using a Cationic Surfactant BEN KWOK-WAI MAN,† M I C H A E L H O N - W A H L A M , * ,† PAUL K. S. LAM,† RUDOLF S. S. WU,† AND GLEN SHAW‡ Centre for Coastal Pollution and Conservation, Department of Biology & Chemistry, City University of Hong Kong, Hong Kong SAR, China, and National Research Centre for Environmental Toxicology, The University of Queensland, Queensland, Australia

A new cloud-point extraction and preconcentration method using a cationic surfactant, Aliquat-336 (tricaprylylmethylammonium chloride), has been developed for the determination of cyanobacterial toxins, microcystins, in natural waters. Sodium sulfate was used to induce phase separation at 25 °C. The phase behavior of Aliquat-336 with respect to concentration of Na2SO4 was studied. The cloud-point system revealed a very high phase volume ratio compared to other established systems of nonionic, anionic, and cationic surfactants. At pH 6-7, it showed an outstanding selectivity in analyte extraction for anionic species. Only MC-LR and MC-YR, which are known to be predominantly anionic, were extracted (with averaged recoveries of 113.9 ( 9% and 87.1 ( 7%, respectively). MCRR, which is likely to be amphoteric at the above pH range, was not detectable in the extract. Coupled to HPLC/ UV separation and detection, the cloud-point extraction method (with 2.5 mM Aliquat-336 and 75 mM Na2SO4 at 25 °C) offered detection limits of 150 ( 7 and 470 ( 72 pg/ mL for MC-LR and MC-YR, respectively, in 25 mL of deionized water. Repeatability of the method was 7.6% for MC-LR and 7.3% for MC-YR. The cloud-point extraction process can be completed within 10-15 min with no cleanup steps required. Applicability of the new method to the determination of microcystins in real samples was demonstrated using natural surface waters collected from a local river and a local duck pond spiked with realistic concentrations of microcystins. Effects of salinity and organic matter (TOC) content in the water sample on the extraction efficiency were also studied.

Introduction Application of the cloud-point phase separation behavior of certain amphiphiles, surfactants with both hydrophilic and * To whom correspondence should be addressed at the Centre of Coastal Pollution and Conservation, Department of Biology & Chemistry, City University of Hong Kong, Tat Chee Ave., Kowloon, Hong Kong. Phone: (852)-2788-7329. Fax: (852)-2788-7406. E-mail: [email protected]. † City University of Hong Kong. ‡ The University of Queensland. 10.1021/es020620v CCC: $22.00 Published on Web 08/13/2002

 2002 American Chemical Society

hydrophobic components in their molecular structures, in aqueous media for the analytical determination of trace organic analytes has aroused growing attention in recent years (1-3). Cloud-point phase separation is the phenomenon of which aqueous solutions of some surfactants undergo phase separation upon certain conditions such as heating or cooling to above or below certain threshold temperature (46) and the addition of salts or acids (7, 8). The result is the separation of two apparently immiscible phases: a surfactantrich phase and an aqueous phase with concentration of surfactant close to the critical micellar concentration (CMC) (1, 4). It has been demonstrated that the surfactant-rich phase, thus separated under the cloud-point conditions, is able to extract and preconcentrate a wide range of organic compounds from the aqueous phase. The advantages of such cloud-point extraction techniques include the following: (a) fast extraction time; (b) high preconcentration factor because of the small volume and usually high analyte affinity for the surfactant-rich phase; and (c) removal of the need to use toxic and environmental unfriendly organic solvents. Up to now, nonionic surfactants, such as Triton X (polyoxyethylene(n)-octylphenyl ether) (9-12), PONPE (polyoxyethylene-(n)nonylphenyl ether) (8, 13, 14), Genapol X (oligoethylene glycol monoalkyl ether), and Brij (polyoxyethylene-10-cetyl ether) (15-17), are the most widely used amphiphiles for cloudpoint extraction. These surfactants have been successfully applied to extract polycyclic aromatic hydrocarbons (PAHs); polychlorinated compounds such as polychlorinated biphenyls (PCBs), polychlorinated dibenzofurans (PCDFs), and dibenzo-p-dioxins (PCDDs); synthetic pesticides; hydroxyaromatic compounds; vitamins; hydrophobic membrane proteins and pharmaceuticals from natural waters, soils and sediments as well as complex biological fluids. Recently, the cloud-point extraction methodology has been extended to utilizing zwitterionic surfactants (13, 18) and anionic surfactants (8). On the other hand, although cloud-point phenomena of cationic surfactants have also been studied (19), application of cationic surfactants in cloud-point extraction is scarce (20). In most cases, zwitterionic or anionic/cationic surfactants were used as cloud-point extractants to overcome the strong UV/visible absorbing interference of nonionic surfactants (8, 21). Relatively little attention has been paid to the possible contribution of the charged headgroups of these surfactants to analyte extraction. From simple consideration of electrostatic interaction, the charged headgroups of cationic amphiphiles should have significant affinity for anionic species in the aqueous media. Thus, it is envisioned that cationic surfactants should be able to extract those anionic analytes, which possess a certain degree of lipid solubility, into the surfactant-rich phase brought about by cloud-point separation. Biomolecules, such as proteins and polypeptides, with both aqueous and lipid solubility are some of the possible candidates. The objective of this work is to explore the feasibility of cloud-point extraction and preconcentration of a group of relatively hydrophilic polypeptide algal toxins, microcystins, from aqueous media using a cationic surfactant, Aliquat-336 (tricaprylylmethylammonium chloride). The algal toxins are a group of cyclic heptapeptide liver toxins produced by freshwater cyanobacteria (blue-green algae) (22-25). The general structure of microcystins is cyclo(-D-Ala-L-R1-Derythro-β-methylisoAsp-L-R2-Adda-D-iso-Glu-N-methyldehydroAla), where R1 and R2 represent two variable L-amino acids and Adda stands for 3-amino-9-methoxy-2,6,8-trimethyl-10-phenyldeca-4,6-dienoic acid. There are already over 60 microcystin (MC) variants known, and MC-LR, -RR, and VOL. 36, NO. 18, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3985

-YR (Figure 1) are the 3 most abundant variants in most natural algal blooms (25). Some of these microcystin variants are known to be potent tumor promoters, as they inhibit activities of serine/threonine protein phosphatases 1 (PP-1) and 2A (PP-2A) (26). Because of this, there is a growing demand for the development of more reliable, rapid, and sensitive analytical determination methods for microcystins in waters (27-31).

Experimental Section Chemicals and Reagents. Standards of microcystin-LR (MCLR), microcystin-YR (MC-YR), and microcystin-RR (MC-RR) were obtained from Calbiochem (La Jolla, CA). HPLC-grade acetonitrile was from Riedel-de Hae¨n (RDH) (Germany). All other chemicals were of analytical reagent grade and were used as received. Cationic surfactant Aliquat-336 (A336) was obtained from Acros. Sodium sulfate was obtained from Aldrich. Trifluoroacetic acid (TFA) was from RDH. Milli-Q water (18 MΩ) used in all experiments was obtained through a Milli-Q (Millipore, Bedford, MA) water purification system. Instrumentation. A Waters 600E HPLC with programmable multiwavelength detector (Waters 490E) and a Beckman Ultrasphere ODS column (5 µm, 250 × 4.6 mm i.d.) (Fullerton, CA) were used for microcystin determination. A gradient elution involving two mobile phases: mobile phase A of 30% acetonitrile and 0.05% trifluoroacetic acid in an aqueous solution and mobile phase B of 100% acetonitrile, was adopted for the HPLC separation of the microcystins. The detection wavelength was set at 238 nm. A Branson Ultrasonic bath and a Thermolyne type 37600 Vortex Mixer were used for mixing the aqueous sample with the surfactant. A Hettich universal centrifuge was used for separation of the surfactant-rich phase. Salinity of natural water samples was determined by an Atago (Japan) hand refractometer. Total organic carbon (TOC) content in natural water samples was determined by a Shimadzu TOC-500 analyzer (Japan). Sampling of Natural Waters. A natural river water sample and a pond water sample were collected from a local intertidal river, known as Lam Tsuen River, and a duck pond in the Mai Po Nature Reserve (where microcystin-producing cyanobacterial blooms are occasionally observed), Hong Kong, respectively, in March 2001. These surface water samples were spiked with known amounts of microcystins and were allowed to age for 24 h at 4 °C before cloud-point extraction. Evaluation of the Phase Behavior of Surfactant. The phase behavior of Aliquat-336 was studied by addition of an appropriate amount of Na2SO4 and Aliquat-336 in glass tubes and made up to 50 mL with Milli-Q water. After 48 h of standing at 25 °C, the tubes were inspected for phase separation. The phase diagram was obtained from these observations as a function of surfactant and Na2SO4 concentrations. Phase volume ratios of the surfactant-rich phase were determined by careful isolation of the phase by a graduated syringe. Cloud-Point Extraction Procedure. In a typical cloudpoint extraction experiment, a known amount of Aliquat336 and Na2SO4 was added to the aqueous sample (25 mL) in a glass centrifuge tube. The mixture was mixed by ultrasonication for 15 min and then vortexed for 1 min. Separation of the two phases was achieved by centrifugation for 10 min at 4000 rpm. The surfactant-rich phase was isolated and mixed with acetonitrile (100 µL) in order to reduce sample viscosity. Finally, a known volume of the diluted mixture was directly injected into the HPLC system without any cleanup step. Standard addition of known concentrations of toxins into the surfactant-rich extract was adopted for the quantitative determination of toxins. 3986

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 18, 2002

FIGURE 1. Chemical structures of the three common microcystin variants: MC-LR, MC-YR, and MC-RR.

Results and Discussion Three common microcystin variants, namely, MC-LR, MCYR, and MC-RR, were used throughout the method development of the cloud-point extraction. Besides their predominance in most of the natural occurring blooms, they are also the only commercially available microcystin standards. Toxin quantification and toxicity evaluation of other microcystin variants are usually undertaken by referring to MC-LR (27). Reverse-phase HPLC with an UV absorption detector was used for quantitative determination of these toxins. As the structure of microcystins contains numerous ionizable functionalities (Figure 1), the overall charge on the toxin molecule is pH-dependent. For example, both MC-LR and MC-YR contain two carboxylic groups and one guanidine group. It has been demonstrated that the cationic MC-LR species, [(COOH)2(NH2+)], is produced at pH 2.19 (32, 33). The additional acidic phenol group at the R1 position of MC-YR is not expected to change the ionization behavior of the variant much from that of MC-LR. Thus, both MC-LR and MC-YR are negatively charged at the pH range of natural waters (pH 6-8). The situation with MC-RR is more complicated as it possesses two carboxylic and two guanidine functionalities. Although the exact pH dependency of the overall charge status of MC-RR has not been reported, our results suggest that it is most likely to be amphoteric at neutral pH. Hydrophilicity of microcystins is a more controversial issue as different measurement methods give rise to different conclusions. Rivasseau et al. (33) reported retention factors, log kw, for MC-LR, MC-YR, and MC-RR to be around 4 using reverse-phase chromatography. As the retention factor is an indirect measurement of the octanol/ water partition coefficient, Kow, Rivasseau et al. concluded that microcystins are rather hydrophobic, despite their amphoteric or anionic character at neutral pH. On the other hand, Maagd et al. (32) measured the octanol/water distribution ratio, Dow, of MC-LR by direct equilibrium method and reported the log Dow of MC-LR to range from 2.18 at pH 1 to -1.76 at pH 10. As Dow is considered a more suitable measurement of hydrophobicity for microcystins than Kow since it takes into account all the species present at the given pH that contribute to the partitioning between octanol and water, Maagd et al. concluded that microcystins are quite hydrophilic at neutral and basic pH. Our preliminary results obtained from attempted cloud-point extraction of microcystins by nonionic surfactant such as Triton X-114 also indicate the hydrophilicity of the toxins. Even with a toxin concentration up to 1 µg/mL in the aqueous phase, no detectable microcystins were extracted into the nonionic surfactant-rich phase.

FIGURE 2. Phase diagram ([Na2SO4] vs [surfactant]) of Aliquat-336 at 25 °C. L, homogeneous region; L - L, two-phase region.

TABLE 1. Phase Volume Ratio of Aliquat-336/Na2SO4 Cloud-Point Mixture at Various Surfactant and Na2SO4 Concentrations at 25 °C [Aliquat-336] (mM)

[Na2SO4] (mM)

phase volume ratio

1.0 1.0 1.0 2.5 2.5 2.5 7.5 7.5 7.5

10 50 75 10 50 75 10 50 75

(2.9 ( 0.4) × 103 (2.5 ( 0.3) × 103 (2.2 ( 0.3) × 103 (1.1 ( 0.2) × 103 (9.5 ( 1.4) × 102 (8.3 ( 1.2) × 102 (2.9 ( 0.5) × 102 (3.0 ( 0.5) × 102 (2.8 ( 0.4) × 102

Cloud-point formation in ionic surfactants is commonly induced by increasing the ionic strength of the aqueous media, i.e., “salting-out” (19). Jin et al. (20) have tried to use sodium chloride to achieve clouding of Cetrimide but were not able to bring about clear phase separation unless a cosurfactant, 1-octanol, was also added. For Aliquat-336 (tricaprylylmethylammonium chloride), cloud-point phase separation is readily accomplished by addition of sodium sulfate. The ability of sulfate to induce phase separation in an aqueous solution of nonionic surfactants has been well documented (10, 34, 35). Figure 2 shows the phase diagram of concentration of Na2SO4 vs Aliquat-336 concentration (up to 15 mM) at 25 °C. Throughout the range of SO42- and surfactant concentrations, only two regions: homogeneous (L) and coexistence of two separated isotropic phases (L L), were observed. No precipitation of surfactant was observed up to 0.6 M Na2SO4. Table 1 shows the phase volume ratio of the cloud-point mixture, which is defined as the ratio of the volume of aqueous solution to that of the surfactant-rich phase, at various surfactant and Na2SO4 concentrations. In general, the phase volume ratio is much higher compared to other established cloud-point systems using zwitterionic, anionic, or other cationic surfactants (3, 8, 20) and shows a decreasing trend with increasing concentrations of surfactant and salt. As the preconcentration factor of a cloud-point extraction method is related to the phase volume ratio, the high phase volume ratio of the Aliquat-336/Na2SO4 system implies outstanding preconcentration ability, especially toward analytes with high affinity for the surfactant-rich phase (3). The ability of the Aliquat-336/Na2SO4 cloud-point system to extract microcystins from aqueous media is shown in Figure 3. HPLC analysis of a typical Aliquat-336/Na2SO4 cloudpoint extract from an aqueous solution (unbuffered, pH 6-7) spiked with an equal amount (100 ng/mL) of MC-LR, -YR, and -RR variants shows chromatographic peaks of MC-LR and MC-YR, but no MC-RR. That is, the Aliquat-336/Na2SO4 system is selective toward extraction of MC-LR and MC-YR

FIGURE 3. Chromatograms of microcystin variants: (a) in Aliquat336 (0.63 M)/acetonitrile mixture. Concentration of MC-LR, MC-YR, and MC-RR was 2.0 ng/mL. (b) In cloud-point extracted from aqueous media by Aliquat-336 (2.5 mM)/[Na2SO4] (75 mM) at 25 °C. Volume of aqueous samples was 25 mL, and amount of MC-LR, MC-YR, and MC-RR in the samples was 10 ng/mL each. The surfactant-rich phase was diluted by 100 µL of acetonitrile before injection. Injection volume was 25 µL. only. MC-RR was not detected in the surfactant-rich phase even when the aqueous concentration was increased to 1.0 µg/mL. This discrimination in extraction efficiency is postulated as being due to the charge states of the toxins. While all three microcystin variants are considerably hydrophilic, both MC-LR and MC-YR are anionic at the pH of the aqueous sample and have high affinity for the Aliquat-336-rich phase via ion-pairing with the cationic headgroups of the surfactant. On the other hand, combination of the hydrophilicity and amphoteric properties of MC-RR at the pH of the aqueous sample may render distribution of the variant into the surfactant-rich phase unfavorable. Figure 4 shows the recoveries of MC-LR and MC-YR by Aliquat-336/Na2SO4 cloud-point extraction at various surfactant and salt concentrations. Although the phase volume ratio decreases with increasing salt concentration, an opposite trend was observed in analyte recovery. While this may be attributable to the additional salting-out effect of the Na2SO4 to favor partitioning of the toxins into the surfactant-rich phase, the possibility of sensitization of the HPLC detector (UV absorption) responses of the toxins by the Aliquat-336 surfactant or [SO42-] cannot be ruled out (3). Sensitization of the signal of the VOL. 36, NO. 18, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3987

FIGURE 5. Linearity of Aliquat-336/Na2SO4 cloud-point extraction and preconcentration of MC-LR (9) and MC-YR (2) from spiked aqueous sample (pH ≈6) at 25 °C. Concentration of Aliquat-336 and Na2SO4 was 2.5 and 75 mM, respectively. Volume of aqueous samples was 25 mL, and the surfactant-rich phase was diluted by 100 µL of acetonitrile before injection. Injection volume was 50 µL.

TABLE 2. Comparison of Performance of Representative Determination Methods for Microcystins method detection limit (ng/mL) quantification method

FIGURE 4. Recovery of microcystins by Aliquat-336/Na2SO4 cloudpoint extraction at various surfactant and salt concentrations at 25 °C: (a) MC-LR; (b) MC-YR. Concentration of Aliquat-336: 1.0 mM (9), 2.5 mM (2), 7.5 mM ([). Each data point shows the mean recovery value with error bar representing (1 SD (n ) 3). Concentration of both toxins in the aqueous samples was 10.0 ng/ mL. Volume of aqueous samples was 25 mL, and the surfactant-rich phase was diluted by 100 µL of acetonitrile before injection. Injection volume was 25 µL. toxins also explains the apparent >100% recovery of MC-LR by the cloud-point extraction. As the standard addition method was used in all subsequent analyses of microcystins in spiked natural water samples, the accuracy of determination was not likely to be affected by such sensitization phenomena. The recovery study also revealed the relationship between extraction efficiency and surfactant concentration. At low Aliquat-336 concentration, extraction efficiency increased with surfactant concentration. However, extraction efficiency dropped as the concentration of the surfactant increases. In all subsequent analyses of spiked and real samples, a cloud-point system of 2.5 mM Aliquat-336 and 75 mM Na2SO4 was adopted. The phase volume ratio at such cloud-point extraction conditions was 8.3 × 102, and averaged analyte recoveries (n ) 9) for MC-LR and MC-YR were 113.9 ( 9% and 87.1 ( 7%, respectively. Figure 5 shows the linearity of the Aliquat-336/Na2SO4 cloud-point extraction and preconcentration method for the determination of MC-LR and MC-YR in a spiked aqueous sample. The method shows good linearity up to a toxin concentration of 50 ng/mL in an aqueous sample. Respective method detection limits for the two microcystin variants (n ) 7) are 150 ( 7 pg/mL for MC-LR and 470 ( 72 pg/mL for MC-YR with a sample volume of 25 mL at 25 °C. Under the same cloud-point extraction conditions, the method achieves 3988

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 18, 2002

MC-LR

ELISA (monoclonal antibody) 0.008 PP2A inhibition assay 0.06 solid-phase extraction, 0.02 HPLC/UV 0.03 HPLC/chemiluminescent 0.02 detection CE/UV detection 0.12 CE/electrospray MS 200 solid-phase microextraction, 7.4 HPLC/UV Aliquat-336 cloud-point 0.15 extraction, HPLC/UV a

MC-YR

MC-RR

ref

0.011 naa 0.02

0.011 na 0.02

(43) (44) (45)

0.03 0.02

0.03 0.02

(33) (46)

0.12 na 7.9

0.12 na 7.1

(47) (48) (42)

0.47

na

na ) not reported.

a repeatability of 7.6% for MC-LR and 7.3% for MC-YR (n ) 6). As the extraction and preconcentration method involves total extraction of toxins from the aqueous phase into the much-condensed surfactant-rich phase, it is envisioned that the detection limits can be further lowered with a larger sample volume. Nevertheless, the present level of sensitivity and repeatability is already comparable to other established chemical determination methods for microcystins. Rivasseau et al. (33) reported a solid-phase extraction HPLC method with detection limits of 30 and 100-200 ng/L for microcystins (MC-LR, -YR, and -RR) in 500 mL of drinking and surface waters, respectively. The typical level of microcystins in moderately contaminated waters is reported to be 0.06-200 µg/L (36, 37). The provisional guideline value for MC-LR in drinking water by WHO is 1.0 µg/L, while a safety limit of 0.1 µg/L has also been suggested taking into consideration chronic effects of the toxins (38). Table 2 compares the performance of the present method with various other representative chemical and biochemical determination methods for microcystins in the literature. As ion-pairing seems to be the major interaction between microcystins and Aliquat-336 in the cloud-point extraction and preconcentration process, it is crucial to find out whether other anionic species in the aqueous media are able to compete with the anionic toxins. Ionic interference is a particularly important issue in environmental analysis as natural waters usually contain significant amounts of mineral salts. The method should be applicable to the analysis of microcystins in saline samples in addition to freshwater. Figure 6 shows the extraction efficiency, in terms of HPLC/

FIGURE 6. Variation of extraction efficiency, in terms of HPLC/UV responses, of the Aliquat-336/Na2SO4 cloud-point extraction of (a) MC-LR and (b) MC-YR, with ionic strength (concentration of NaCl) of the aqueous media at 25 °C. Each data point shows the mean HPLC/UV response with the error bar representing (1 SD (n ) 3). Concentration of Aliquat-336 and Na2SO4 was 2.5 and 75 mM, respectively. Concentration of MC-LR and MC-YR in the spiked aqueous samples was 10.0 ng/mL. Volume of aqueous samples was 20 mL, and the surfactant-rich phase was diluted by 100 µL of acetonitrile before injection. Injection volume was 25 µL. UV responses, of the Aliquat-336/Na2SO4 cloud-point extraction in aqueous media of different ionic strengths ranging from freshwater to 0.6 M NaCl solution (equivalent to the salinity of seawaters). Our results reveal that chloride ion, the most abundant anion in natural saline waters, only causes slight to moderate reduction in the efficiency of the cloudpoint extraction. In addition, no significant change in phase volume ratio was observed with the addition of NaCl to the aqueous samples. The mean extraction efficiency for MC-LR and MC-YR at 0.6 M NaCl was 66.1% and 88.0%, respectively, of that in freshwater (i.e., no NaCl). Thus, sensitivity of the Aliquat-336/Na2SO4 cloud-point extraction and determination of microcystins should still be maintained even with samples with ionic strength as high as seawater. To examine the performance of the Aliquat-336/Na2SO4 cloud-point extraction and preconcentration method for the determination of microcystins in natural waters, the method was applied to local river- and pond-water samples spiked with “realistic” concentrations of MC-LR and MC-YR. Figure 7 shows the HPLC/UV responses of the cloud-point extracts of the two surface water samples at two spiked levels of MCLR and MC-YR. The overall cloud-point extraction process took only 10-15 min with no need for cleanup steps after phase separation. Table 3 summarizes important physicochemical parameters of these natural waters as well as the

FIGURE 7. HPLC/UV responses of Aliquat-336/Na2SO4 cloud-point extracts from local natural waters: (a) river water from Lam Tsuen River, Taipo, Hong Kong; (b) pond water from a duck pond in Mai Po Nature Reserve, Hong Kong. Spike levels for both MC-LR and MC-YR were 1.0 and 5.0 ng/mL. Each data point shows the mean HPLC/UV response with the error bar representing (1 SD (n ) 3). Concentration of Aliquat-336 and Na2SO4 was 2.5 and 75 mM, respectively. Volume of aqueous samples was 25 mL, and the surfactant-rich phase was diluted by 100 µL of acetonitrile before injection. Injection volume was 50 µL.

TABLE 3. Physicochemical Parameters of Natural Water Samples and Analyte Recovery of the Cloud-Point Extraction river water from Lam Tsuen River, TaiPo, Hong Kong pH salinity (ppt)a TOC (ppm) recovery (%)b MC-LR MC-YR a

6.5 0.5 4.7 85.0 ( 17 96.6 ( 19

Parts-per-thousand.

b

freshwater from duck pond in Mai Po Nature Reserve, Hong Kong 8.4 17.0 23.6 42.2 ( 7 39.9 ( 12

Averaged recovery, n ) 6.

performance of the cloud-point extraction. The lower toxin recovery from the pond water sample is ascribable to a combination of higher salinity, higher total organic carbon (TOC) content, and the pH of the sample matrix. The effect of high pH on the efficiency of the cloud-point extraction method may not be significant, as it does not cause any further change in the overall charge of the toxins. On the other hand, numerous studies have revealed strong affinity of trace organic species in natural waters for organic matter (39-41). Such adsorption between analytes and organic matters can VOL. 36, NO. 18, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3989

cause lowering of analyte recovery (41). Our previous study also demonstrated the association of microcystins to natural suspended particles, causing the discrepancy in the intraand extracellular profiles of microcystins in natural bloom waters (42). Indeed, MC-YR was found to possess greater ability to adsorb onto polar surfaces of suspended materials than MC-LR, probably because of the additional polar phenolic functional group on the R1 position. This may account for the greater drop in recovery of MC-YR in samples with higher TOC content. The present study demonstrated the ability of the cationic surfactant, Aliquat-336, to extract and preconcentrate microcystins from aqueous media via cloud-point extraction. Cloud-point phase separation was conveniently induced by Na2SO4 at room temperature. The cloud-point extraction process was characterized by its high efficiency, with an outstanding degree of preconcentration as well as generally good analyte recovery, and high selectivity toward the charge states of its analytes. At pH 6-7, the cloud-point extraction is very efficient in preconcentrating anionic microcystin variants, such as MC-LR and MC-YR, but shows negligible affinity for MC-RR, which is likely to be amphoteric at the pH. Ion-pairing between the anionic toxin species and the cationic headgroup of surfactant molecules followed by the extraction of the much more hydrophobic ion-pairs into the surfactant-rich phase is the most probable mechanism for the cloud-point extraction and preconcentration of the hydrophilic toxins from the aqueous phase. Coupled with HPLC/UV, the Aliquat-336/Na2SO4 cloud-point extraction system offers a rapid and reliable method with good sensitivity for the determination of MC-LR and MC-YR with relatively small sample volume. Despite the ionic nature of interaction between the analytes and the cationic surfactant, the method was not greatly affected by ionic interference. On the other hand, application of the method to spiked natural water samples revealed that organic matter in the sample affects analyte recovery significantly, especially for MC-YR. Nevertheless, such interference can be overcome by standard addition during quantitative analysis, or estimated by the use of suitable surrogates. Work is in progress to extend such an Aliquat-336/Na2SO4 cloud-point system as well as systems with other cationic surfactants to other applications involving determination of anionic, hydrophilic macromolecules in natural waters and biological fluids.

Acknowledgments The work described in this paper was funded by a Central Allocation Grant (Project 8730011) and an Earmarked Research Grant (CityU 1103/00M) from the Research Grants Council of the Hong Kong Special Administrative Region, China.

Literature Cited (1) Gullickson, N. D.; Scamehom, J. F.; Harwell, J. H. In SurfactantBased Separation Processes; Schamehom, J. F., Harwell, J. H., Eds.; Marcel Dekker: New York, 1989; pp 139-151. (2) Hinze, W. L.; Pramauro, E. CRC Crit. Rev. Anal. Chem. 1993, 24, 133. (3) Carabias-Martı´nez, R.; Rodrı´guez-Gonzalo, E.; Moreno-Cordero, B.; Pe´rez-Pavo´n, J. L.; Garcı´a-Pinto, C.; Ferna´ndez Laespada, E. J. Chromatogr. A 2000, 902, 251 and references cited therein. (4) Watanabe, H. In Solution Behaviour of Surfactants; Mittal, K. L., Fendler, E. F., Eds.; Plenum Press: New York, 1982; Vol. 2, p 1305. (5) Corti, M.; Minero, C. J. Phys. Chem. 1984, 88, 309. (6) Blankschtein, D.; Thurston, G. M.; Benedek, G. B. J. Chem. Phys. 1986, 85, 7268. (7) Schott, H. J. Colloid Interface Sci. 1997, 192, 458. (8) Casero, I.; Sicilia, D.; Rubio, S.; Pe´rez-Bendito, D. Anal. Chem. 1999, 71, 45. (9) Sirimanne, S. R.; Barr, J. R.; Patterson, D. G.; Ma, L. Anal. Chem. 1996, 68, 1556. 3990

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 18, 2002

(10) Fang, Q.; Yeung, H. W.; Leung, H. W.; Huie, C. W. J. Chromatogr. A 2000, 904, 47. (11) Garcı´a-Pinto, C.; Pe´rez-Pavo´n, J. L.; Moreno-Cordero, B. Anal. Chem. 1995, 67, 2696. (12) Wu, Y. C.; Huang, S. D. Anal. Chim. Acta 1998, 373, 197. (13) Saitoh, T.; Hinze, W. L. Anal. Chem. 1991, 63, 2520. (14) Silva, M. F.; Fernandez, L. P.; Olsina, R. A. Analyst 1998, 123, 1803. (15) Eiguren-Ferna´ndez, A.; Sosa-Ferrera, Z.; Santana-Rodrı´guez, J. J. Anal. Chim. Acta 1998, 358, 145. (16) Sirimanne, S. R.; Patterson, D. G.; Ma, L.; Justice, J. B. J. Chromatogr. B 1998, 716, 129. (17) Eiguren-Ferna´ndez, A.; Sosa-Ferrera, Z.; Santana-Rodrı´guez, J. J. Analyst 1999, 124, 487. (18) Saitoh, T.; Hinze, W. L. Talanta 1995, 42, 119. (19) Rosen, M. J. Surfactants and Interfacial Phenomena; Wiley: New York, 1989. (20) Jin, X.; Zhu, M.; Conte, E. D. Anal.Chem. 1999, 71, 514. (21) Cordero, B. M.; Pavon, J. P. L.; Pinto, C. G.; Lasepada, M. E. F. Talanta 1993, 40, 1703. (22) Carmichael, W. W.; Falconer, I. R. In Algal Toxins in Seafood and Drinking Water; Falconer, I. R., Ed.; Academic Press: London, 1993; pp 197-209. (23) Codd, G. A. Wat. Sci. Technol. 1995, 32, 149. (24) Duy, T. N.; Lam, P. K. S.; Shaw, G. R.; Connell, D. W. Rev. Environ. Contam. Toxicol. 2000, 163, 113. (25) Harada, K.-I. In Toxic Microcystis; Watanabe, M. F., Harada, K.-I., Carmichael, W. W., Eds.; CRC Press: Boca Raton, FL, 1996; p 103. (26) Yu, S. Z. In Primary Liver Cancer; Tang, Z. Y., Wu, M. C., Xia, S. S., Eds.; Springer-Verlag: Berlin, 1989; pp 30-37. (27) Meriluoto, J. Anal. Chim. Acta 1997, 352, 277 and references cited therein. (28) Chu, F. S.; Huang, X.; Wei, R. D. J. Assoc. Off. Anal. Chem. 1990, 73, 451. (29) Nagata, S.; Tsutsumi, T.; Yoshida, F.; Ueno, Y.; Watanabe, M. F. J. Assoc. Off. Anal. Chem. 1997, 80, 408. (30) Holmes, C. F. B. Toxicon 1991, 29, 469. (31) Lambert, T. W.; Boland, M. P.; Holmes, C. F. B.; Hrudey, S. E. Environ. Sci. Technol. 1994, 28, 753. (32) Maagd, P. G.; Hendriks, A. J.; Seinen, W.; Sijm, D. T. H. M. Wat. Res. 1999, 33, 677. (33) Rivasseau, C.; Martins, S.; Hennion, M. C. J. Chromatogr. A 1998, 799, 155. (34) Parish, C. R.; Classon, B. J.; Tsagaratos, J.; Walker, I. D.; Kirszbaum, L.; McKenzie, I. F. C. Anal. Biochem. 1986, 156, 495. (35) Horvath, W. J.; Huie, C. W. Talanta 1993, 40, 1385. (36) Ueno, Y.; Nagata, S.; Tsutsumi, T.; Hasegawa, A.; Watanabe, M. F.; Park, H.-D.; Chen, G.-C.; Chen, G.; Yu, S.-Z. Carcinogenesis 1996, 17, 1317. (37) McDermott, C. M.; Feola, R.; Plude, J. Toxicon 1995, 33, 1433. (38) Ressom, R.; Soong, F. S.; Fitzgerald, J.; Turczynowicz, L.; El Saadi, O.; Roder, D.; Maynard, T.; Falconer, I. Health effects of toxic cyanobacteria (blue-green algae), Report to the Environmental Health Standing Committee of the National Health and Medical Research Council, University of Adelaide and South Australian Health Commission, Adelaide, 1993; p 92. (39) Landrum, P. F.; Nihart, S. R.; Eadie, B. J.; Gardner, W. S. Environ. Sci. Technol. 1984, 18, 187. (40) Schomburg, C. J.; Glotfelty, D. E.; Seiber, J. N. Environ. Sci. Technol. 1991, 25, 155. (41) Johnson, W. E.; Fendinger, N. J.; Plimmer, J. R. Anal. Chem. 1991, 63, 1510. (42) Poon, K. F.; Lam, M. H. W.; Lam, P. K. S.; Wong, B. S. F. Environ. Toxicol. Chem. 2001, 20, 1648. (43) Seck, A.; Eikenberg, A.; Weller, M. G.; Niessner, R. Anal. Chim. Acta 2001, 441, 1. (44) Serres, M. H.; Fladmark, K. E.; Døskeland, S. O. Toxicon 2000, 38, 347. (45) Lee, H. S.; Jeong, C. K.; Lee, H. M.; Choi, S. J.; Do, K. S.; Kim, K.; Kim, Y. H. J. Chromatogr. A 1999, 848, 179. (46) Murata, H.; Shoji, H.; Oshikata, M.; Harada, K.-I.; Suzuki, M.; Kondo, F.; Goto, H. J. Chromatogr. A 1995, 693, 263. (47) Onyewuenyi, N.; Hawkins, P. J. Chromatogr. A 1996, 749, 271. (48) Bateman, K. P.; Thibault, P.; Douglas, D. J.; White, R. L. J. Chromatogr. A 1995, 712, 253.

Received for review February 27, 2002. Revised manuscript received June 24, 2002. Accepted July 2, 2002. ES020620V