Elyns Journal of Pharmaceutical Research

Development of a Solid Phase Extraction Method to Extract Gonyautoxins from Urine

Download PDF

Published On January 15, 2015

Development of a Solid Phase Extraction Method to Extract Gonyautoxins from Urine

Padmanabhan Eangoor, Amruta Indapurkar and Jennifer S. Knaack*
Department of Pharmaceutical Sciences, Mercer University, 3001 Mercer University Dr., Atlanta, GA 30341, USA

*Corresponding author: Jennifer S. Knaack, Department of Pharmaceutical Sciences, Mercer University, 3001 Mercer University Dr., Atlanta, GA 30341,Phone: 678-547-6737; Email: knaack_js@mercer.edu

Citation: Eangoor P, Indapurkar A, Knaack JS (2015) Development of a Solid Phase Extraction Method to Extract Gonyautoxins from Urine. Ely J Pharm Res 1(1): 101.



Saxitoxin (STX), Neosaxitoxin (NEO) and Gonyautoxins (GTXs) I-IV form a carbamate subgroup of hydrophilic Paralytic Shellfish Toxins (PSTs) that can be found in shellfish that have fed upon red tide algae. Analytical methods to detect these toxins have been developed primarily for the analysis of contaminated shellfish. Diagnostic tests for measuring exposure to these toxins is limited to a single urinalysis method for STX and NEO and is confirmed by analysis of unconsumed shellfish. Validated diagnostic methods to measure exposure to GTXs from urine do not exist and are needed for accurate diagnosis of paralytic shellfish poisoning. Here we describe solid phase extraction (SPE) approaches to extract GTXs from urine in an effort to develop an analytical diagnostic method for GTX exposure. Three SPE sorbents have been explored in this study: mixedmode strong cationic exchange.weak cationic exchange, and mixed-mode weak anionic exchange. No toxins were recovered from weak anion exchange extractions. Weak cationic exchange shows higher recoveries for GTXs than strong cationic exchange and may be useful for measuring high-level exposures. However, extraction efficiencies are low due to urine matrix effects on the extraction and future improvements on the method are needed for measuring low-level exposures.

Keywords: Gonyautoxin; Urine; Solid Phase Extraction; Ion-Exchange; HPLC-MS/MS; Paralytic Shellfish Poison. 

Top ↑


HPLC-MS/MS: High performance liquid chromatography coupled to tandem mass spectrometry; SPE: Solid phase extraction; PSP: Paralytic shellfish poisoning; PST: Paralytic shellfish toxin; GTX: Gonyautoxin; STX: Saxitoxin; NEO: Neosaxitoxin.


Gonyautoxins (GTXs) I-IV, saxitoxin (STX), and neosaxitoxin (NEO) form a carbamate group of paralytic shellfish toxins (PSTs) [1] produced by dinoflagellates during algal blooms [2]. The dinoflagellates are fed upon by filter feeding shellfish like mussels, clams, oysters, and scallops. Human exposure to these toxins occurs upon ingestion of contaminated shellfish resulting in paralytic shellfish poisoning (PSP). PSP is characterized by oral paresthesia, asthenia, disthonia, ataxia, dyspnea, hypotension, tachycardia, vomiting and muscular weakness which can be fatal [3]. To date,many human exposures have been reported, but very few have been discussed in peer-reviewed literature [4-7]. The mouse bioassay, considered the gold standard method for PST analysis [8], has been traditionally used to detect and quantify these toxins in shellfish tissue samples. Due to ethical considerations for animal use in toxicity studies [9], and to develop a more reliable analytical method, biochemical and chemical analytical alternatives were explored. Biochemical methods include immunoassays like enzyme-linked immunoassays [10] and chemical analyses include solid phase extraction (SPE) followed by pre-column [11] or postcolumn postcolumn oxidation [12,13] coupled to high performance liquid chromatography (HPLC) with a fluorescence detector, HPLC-UV detector or SPE followed by HPLC-tandem mass spectrometry (HPLC-MS/MS). These analytical methods are currently used to extract PSTs from shellfish tissues and SPE-HPLC-MS/MS has been reported for the analysis of STX and NEO in human urine [14]. The major route of excretion of PSTs in humans is through urine [15]. The lack of urinary diagnostic methods for measuring exposure to carbamate PSTs, coupled with the difficulty of collecting urine in a timely manner [15], have been key limiting factors for diagnosing exposure to carbamate PSTs. Recently, a method was published to effectively extract STX and NEO from urine [14], but was unsuccessful in extracting GTXs (unpublished data) which can comprise a large portion of the toxins present in contaminated shellfish and can be present in shellfish to the exclusion of STX and NEO [16].

The chemistry of GTXs is unique compared to other carbamate toxins in that GTXs can exist as zwitterionic species in certain pH ranges. They exist as positively charged species at a pH below 6 and as negatively charged species at pHs above 4. SPE of ionic substances from urine is usually accomplished by ion-exchange.The ability to exist in both positively charged and negatively charged states potentially allows extraction of GTXs using cationic or anionic exchange extraction principles. The amine groups on GTXs have a pKa of approximately 8.5. While using strong cationic exchange cartridges, GTXs behave as ionic compounds at a pHs below 6.5 and non-ionic compounds at a pHs above 10.5. When using weak cationic exchange SPE cartridges, GTXs are maintained ionic all times whereas the pH condition of the sorbent material is changed to make it ionic and non-ionic as necessary. Weak anionic exchange SPE cartridges are used to target the sulfate groups on the toxins. Sulfates have a pKa of approximately 2 which allows them to remain ionic at pHs above 4 and the pH conditions of the sorbent are changed to make it ionic and non-ionic as necessary. Here, we describe the use of these ion exchange principles to extract GTXs from both urine and water matrices.

Top ↑

Materials and Methods 

HPLC-MS-grade acetonitrile and water were purchased from Med Supply Partners (Atlanta, GA, USA). Formic acid and ammonium formate were also purchased from Med Supply Partners (Atlanta, GA, USA). Ammonium hydroxide was purchased from Fisher Scientific (Pittsburg, PA, USA).HPLC-grade methanol, sodium phosphate dibasic heptahydrate and sodium phosphate monobasic monohydrate were purchased from Sigma-Aldrich (St. Louis, MO, USA). Certified calibration solutions for GTX1,4 and GTX2,3 were obtained from the Institute for Marine Biosciences,National Research Council Canada (Ottawa, Ontario, Canada).Pooled human urine was obtained from Lee Biosolutions (St.Louis, MO, USA).


A 12-port SupelcoVisiprep vacuum manifold (Sigma-Aldrich, St. Louis, MO, USA) was used for solid phase extractions. Mixed-mode Oasis MCX 3cc vac cartridges, 60 mg sorbent, 30 μm particle size (Waters, Milford, MA, USA)were used for strong cationic exchange experiments. Mixed-mode Oasis WAX 3cc vac cartridges, 60 mg sorbent, 30 μm particle size (Waters, Milford, MA, USA) were used for weak anionic exchange experiments. Bakerbond carboxylic acid cartridges, 3 cc capacity, 500 mg sorbent size (Avantor performance materials, Center valley, PA, USA) were used for weak cationic exchange experiments. A Genevac EZ-2 plus vacuum evaporator (SP scientific, Stone ridge, NY, USA) was used to evaporate extracted fractions to dryness. A 1210 Infinity HPLC system (Agilent technologies, Santa clara, CA, USA) with a binary pump, auto sampler, and a temperaturecontrolled column chamber was used with an XBridge BEH Amide hydrophilic liquid interaction chromatography (HILIC)[17-20] column with particle size 2.5 μm and column dimensions 2.1 x 150 mm (Waters, Milford, MA, USA) for the separation of toxins. Separated toxins were detected using a 6410 triple quadrupole mass spectrometer(Agilent technologies, Santa clara, CA, USA)operated under multiple reaction monitoring (MRM) mode [21].

Buffers and solutions
HPLC mobile phases were as follows: the aqueous phase was composed of 2mM ammonium formate buffer, pH 3.5 (phase A) and the organic phase were composed of 0.1% formic acid in acetonitrile (phase B). SPE was performed using liquid chromatography-mass spectrometry (LCMS)-grade water which was basified to pH 6.4 using 0.01% ammonium hydroxide to make it pH 6.4 water. A 500 μL of 5000 ng/mL toxin mix stock solution was made by mixing 76 μL of GTX 1 & 4 (32.9 μg/mL) and 40 μL of GTX 2 & 3(62.3 μg/mL )with water to a total volume of 500 μL. Reconstitution solution was made by mixing 80 parts of acetonitrile with 20 parts of water and adding 0.1 % w/v of formic acid to the final volume .

Sample preparation methods
Toxin-spiked urine specimens were prepared by spiking 10 μL of the toxin mix stock solution into 490 μL of pooled human urine to a final concentration of 100 ng/mL of GTX 1,4 and GTX 2,3.Spiked water samples were prepared by adding 10 μL of toxin mix stock solution to 490 μL of LCMS water so that the concentration of each toxin was100 ng/mL. Unextracted positive controls were prepared by adding a 10 μL aliquot of stock solution to 490 μL of reconstitution buffer. This sample was dried and reconstituted again with 100 μL of reconstitution buffer. For strong cationic exchange experiments, 10 μL of the mixed toxin stock solution was added to 490 μL of 5% ammonium hydroxide in water which was then dried completely and reconstituted in 100 μL of reconstitution solution. This sample served as an “unextracted toxins in base” sample. The spiked urine and spiked water samples were then introduced ontoSPE cartridges of different chemistries and suitable fractions were collected as discussed below (table 1). The collected samples were dried in GeneVac vacuum evaporator and reconstituted with 100 μL of reconstitution solution to make a final concentration of 500ng/ mL of GTX1,4 and GTX2,3.

Protocols for ion-exchange extractions
For strong cationic exchange and weak anionic exchange experiments, the same protocol was followed. However, for strong cationic exchange experiments MCX cartridges were used and for weak anionic exchange experiments, WAX cartridges were used. Extraction methods began with conditioning the sorbent with 500 μL methanol followed by 500 μL water. Prepared samples were then loaded onto the cartridge. The sorbent was then washed with 500 μL of 2% formic acid in water followed by 500 μL of methanol. Finally, the toxins were eluted with a 500 μL aliquot of 5% ammonium hydroxide in methanol. For weak cationic exchange extractions, the sorbent was first conditioned with 500 μL of methanol followed by 1 mL of water at pH 6.4. The spiked sample was then loaded onto the cartridge and the cartridge was then washed with 1 mL of pH 6.4 water followed by 1 mL of methanol. Finally, the toxins were eluted in 2 mL of 5% formic acid in methanol.

HPLC-MS/MS method

A volume of 15 μL of the extracted, dried and reconstituted sample was injected into the HPLC. A gradient method was used as follows: 25% mobile phase A for 2 minutes, gradient to 45%mobile phase A over 6.62 minutes, hold at 45% mobile phase A for 2.21 minutes, hold at 25% mobile phase A for 2.21 minutes. The column temperature was maintained at 30°C and the flow rate was maintained at 0.35 mL/min. Chromatographed samples were detected by electrospray ionization MS/MS in positive mode. The optimized parameters of the mass spectrometer were as follows: drying gas temperature at 350°C, gas flow at 12 L/min, nebulizer pressure at 35 psi, capillary voltage at 4000 V and electron multiplier voltage at 300 V. Quantifier and qualifier ions were optimized using optimizer software and the resultant transitions were used in MRM mode. The following transitions were used for each toxin: GTX 1 (332→236.1), GTX 2 (316 → 298.2), GTX 3 (396.1 →298.2) and GTX 4 (412.2 →314.1). The chromatographic peaks were then quantified using Mass Hunter software.

Calibration curves

Toxin mix stock solution with a concentration of 5000ng/mL of GTX 1,4 and 5000 ng/mL of GTX 2,3 was used to make further dilutions. Serial dilutions from the toxin mix stock solution were performed to make calibrators at the following concentrations: 500ng/mL, 250ng/ mL, 100ng/mL, 50ng/mL and 10ng/mL. Reconstitution solution was used to perform dilutions. These calibrators were then analyzed by HPLC-MS/MS and quantitative analysis software was used to build linear 1/x weighted calibration curves for each toxin (GTX 1,2, 3 and 4, each separately).



Calibration curves
Four calibration curves were generated corresponding to the four GTXs. The coefficients of determinations were 0.9965 for GTX1, 0.9953 for GTX2, 0.9977 for GTX3 and 0.9974 for GTX4. Calibration curves are shown in Figure 1.

Figure 1: Calibration curves of GTX 1, 2, 3 and 4 with concentrations of toxins in ng/mL plotted on x-axis and peak area in counts per second plotted on y-axis.


Strong cationic exchanger (MCX) extraction
For MCX extractions, toxins were eluted in a basic solution which can degrade toxins over time. Therefore, an “unextracted toxins in base” sample was used as positive control and treated as the maximal amount of toxin that could be recovered using this method. Recovery of toxins from MCX extractions was higher for water samples than for urine samples. The peak areas of other extractions were compared to the peak area of ‘unextracted toxin in base’ sample and reported in relative percentages (Figure 2, Table 1). The extractions of toxin spiked water sample were high for GTX1 and GTX4 with recoveries of 79.7 and 54.1%, respectively, whereas the recoveries of GTX2 and GTX3 were less than 1%. A 66.5% reduction in the extraction efficiency was seen for GTX1 whereas an 87.5% reduction in the extraction efficiency was seen for GTX4 in the extraction of toxin spiked urine sample compared to toxin spiked water sample. It is observed that the recovery of GTX2 using MCX cartridges is higher in spiked urine sample compared to the spiked water sample.

Table 1: Table showing percentage recoveries of GTXs 1, 2, 3 and 4 from spiked water sample and spiked urine sample using MCX cartridge, by taking the peak areas in counts per second of unextracted toxins in base as 100 percent


Figure 2: Strong cationic exchange extractions of GTXs 1, 2, 3 and 4. Unextracted toxins in base, Spiked water sample and spiked urine sample extractions of each toxin is shown in terms of peak area on a logarithmic scale of counts per second.


Weak cationic exchanger (WCX) extraction
For WCX extractions, the peak areas of unextracted toxins served as a positive control and were considered as 100% recovery. Weak cationic exchange extractions of spiked water and spiked urine samples were compared to this control to determine the percent recoveries of each toxin (Figure 3, Table 2). The recoveries of all toxins from spiked water sample extractions were considerably high using weak cationic exchange compared to the spiked water sample extraction using strong cationic exchange, and were most notable for GTX2 and GTX3. However, extraction recoveries for all toxins from spiked urine samples decreased drastically with GTX4 being the most affected and GTX2 being the least affected by the urine matrix. Among all toxins, GTX2 was recovered to the highest extent in urine at 57.5% from using a WCX cartridge .


Table 2: Table showing percentage recoveries of GTXs 1, 2, 3 and 4 from spiked water samples and spiked urine samples using WCX cartridges, by taking the peak areas in counts per second of unextracted toxins as 100 percent.


Figure 3: WCX extractions of GTXs 1, 2, 3 and 4. Unextracted toxins and extracted spiked water samples and spiked urine samples are shown in terms of peak area on a logarithmic scale of counts per second.


Weak anionic exchanger (WAX) extraction:
Recovery of toxins from WAX cartridges was 0% for all toxins.

Top ↑


Peak areas significantly vary between the GTXs. The peak areas of unextracted GTX3 and GTX4 are higher than those for GTX1 and GTX2. This is because GTX3 and GTX4 are present at threefold higher concentrations than GTX1 and GTX2 in the reference solutions for GTX1,4 and GTX2,3. Extraction of toxins from spiked water samples were used to determine SPE sorbent efficiency in a pure aqueous solution and also to determine any influence of urine matrix substances in extracting toxins from urine samples. In both weak cationic and strong cationic exchanges, the lower recoveries of all toxins from urine samples can be attributed to the matrix components of urine. These urine components may cause poor binding of toxins to sorbent or might cause premature elution of the toxins from the sorbent. Higher recovery of GTX2 from spiked urine samples compared to spiked water samples using MCX cartridge may be a result of ion-enhancement caused by the urine matrix. The recoveries of GTX1 and GTX4 from spiked water samples using MCX cartridge was higher compared to GTX2 and GTX3 which could be because of the presence of hydroxyl group in their structures allowing them to bind more efficiently to the cartridge. In case of WCX extraction from spiked water, the recoveries of GTX1 and GTX2 were higher than recoveries for GTX3 and GTX4 which may be a result of a similar spatial arrangement of sulfate groups in these molecules. This arrangement can be conducive for binding of amine groups to the sorbent.For weak anionic exchange extractions, a complete absence of toxins in the elutions suggest that the toxins may not have bound to the sorbent but were instead lost in the flow through or during wash steps. Both MCX and WAX sorbents are mixed-mode and contained reversed-phase extraction chemistries that GTXs may have retained by. However, GTXs were not detected in the methanol washes implying that the reversedphase chemistry has no binding capacity for these toxins.

Top ↑

Conclusion and future directions 

Diagnostics for human exposure to GTXs are not currently available, but are required to assess total exposure to PSTs. Such a diagnostic would benefit greatly from an efficient SPE step to prepare samples for downstream analytical analysis. In this study, we found that weak anion exchange sorbents cannot retain GTXs in a urine matrix and that strong cation exchange can retain some GTXs, but not strongly. However, weak cation exchange sorbents provide better retention of GTXs and may be useful for assessing high-level exposures to GTXs. The influence of the urine matrix on extraction efficiency should be further studied to improve retention of toxins on the sorbent. None of the tested sorbents show promise for detection of low-level human exposure to GTXs and so chemical modification of the toxins may be required to improve the extraction efficiency.

Top ↑


MCX and WAX SPE cartridge samples were kindly provided by Waters (Milford, MA, USA).

Top ↑



  1. Zhuo L, Yin Y, Fu W, Qiu B, Lin Z, Yang Y, et al. Determination of paralytic shellfish poisoning toxins by HILIC-MS/MS coupled with dispersive solidphase extraction. Food Chem. 2013; 137 (1-4): 115–21. doi: 10.1016/j. foodchem.2012.10.010.
  2. Tan CT, Lee EJ. Paralytic shellfish poisoning in Singapore. Ann Acad Med Singapore. 1886; 15 (1): 77–79.
  3. García C, Lagos M, Truan D, Lattes K, Véjar O, Chamorro B, et al. Human intoxication with paralytic shellfish toxins: clinical parameters and toxin analysis in plasma and urine. Biol Res. 2005; 38(2-3): 197–205.
  4. Hallegraeff GM. A review of harmful algal blooms and their apparent global increase. Phycologia. 1993; 32: 79–99. http://dx.doi.org/10.2216/ i0031-8884-32-2-79.1.
  5. Yasumoto T, Murata M. Marine toxins. Chem Rev. 1993; 93: 1897–1909.
  6. Falconer IR. Potential impact on human health of toxic cyanobacteria. Phycologia 1996; 35(6s): 6–11. http://dx.doi.org/10.2216/i0031-8884- 35-6S-6.1.
  7. García C, del Carmen Bravo M, Lagos M, Lagos N. Paralytic shellfish poisoning: post-mortem analysis of tissue and body fluid samples from human victims in the Patagonia fjords. Toxicon. 2004; 43(2): 149–158.
  8. Van de Riet JM, Gibbs RS, Chou FW, Muggah PM, Rourke WA, Burns G, et al. Liquid chromatographic post-column oxidation method for analysis of paralytic shellfish toxins in mussels, clams, scallops, and oysters: single-laboratory validation. J AOAC Int. 2009; 92(6): 1690–1704.
  9. Hess P. Requirements for screening and confirmatory methods for the detection and quantification of marine biotoxins in end-product and official control. Anal Bioanal Chem. 2010; 397(5): 1683–1694. doi: 10.1007/s00216-009-3444-y.
  10. Usleber E, Schneider E, Terplan G. Direct enzyme immunoassay in microtitration plate and test strip format for the detection of saxitoxin in shellfish. Lett Appl Microbiol. 1991; 13(6): 275–277. DOI: 10.1111/j.1472-765X.1991.tb00627.x
  11. Stafford RG, Hines HB. Method for the identification of saxitoxin in rat urine. J Chromatogr B Biomed Appl. 1994; 657(1): 119–124.
  12. Oshima Y. Postcolumn derivatization liquid chromatographic method for paralytic shellfish toxins. J AOAC Int. 1995; 78(2): 528–532.
  13. Rourke WA, Murphy CJ, Pitcher G, van de Riet JM, Burns BG, Thomas KM. et al. Rapid postcolumn methodology for determination of paralytic shellfish toxins in shellfish tissue. J AOAC Int 2008; 91(3): 589–597.
  14. Johnson RC, Zhou Y, Statler K, Thomas J, Cox F, Hall S, et al. Quantification of saxitoxin and neosaxitoxin in human urine utilizing isotope dilution tandem mass spectrometry. J Anal Toxicol. 2009; 33(1): 8–14.
  15. Gessner BD, Bell P, Doucette GJ, Moczydlowski E, Poli MA, , Van Dolah F, et al. Hypertension and identification of toxin in human urine and serum following a cluster of mussel-associated paralytic shellfish poisoning outbreaks. Toxicon Off J Int Soc Toxinology. 1997; 35(5): 711–722.
  16. Deeds JR, Petitpas CM, Shue V, White KD, Keafer BA, , McGillicuddy DJ Jr, et al. PSP toxin levels and plankton community composition and abundance in size-fractionated vertical profiles during spring/summer blooms of the toxic dinoflagellate Alexandrium fundyense in the Gulf of Maine and on Georges Bank, 2007, 2008, and 2010: 1. Toxin levels. Deep- Sea Res Part II Top Stud Oceanogr. 2014; 103: 329–349.
  17. Turrell E, Stobo L, Lacaze JP, Piletsky S, Piletska E. Optimization of Hydrophilic Interaction Liquid Chromatography/Mass Spectrometry and Development of Solid-Phase Extraction for the Determination of Paralytic Shellfish Poisoning Toxins. J AOAC Int. 2008; 91(6): 1372–1386.
  18. Diener M, Erler K, Christian B, Luckas B. Application of a new zwitterionic hydrophilic interaction chromatography column for determination of paralytic shellfish poisoning toxins. J Sep Sci. 2007; 30(12): 1821–1826.
  19. Sayfritz SJ, Aasen JA. B, Aune T. Determination of paralytic shellfish poisoning toxins in Norwegian shellfish by liquid chromatography with fluorescence and tandem mass spectrometry detection. Toxicon Off J Int Soc Toxinology. 2008; 52(2): 330-340. doi: 10.1016/j. toxicon.2008.06.001.
  20. Blay P, Hui JPM, Chang J, Melanson JE. Screening for multiple classes of marine biotoxins by liquid chromatography-high-resolution mass spectrometry. Anal Bioanal Chem. 2011; 400(2): 577–585. doi: 10.1007/ s00216-011-4772-2.
  21. McNabb P, Selwood AI, Holland PT, Aasen J, Aune T, Eaglesham G, et al. Multiresidue method for determination of algal toxins in shellfish: single-laboratory validation and interlaboratory study. J AOAC Int. 2005; 88(3): 761–772.


Top ↑

Copyright: © 2015 Jennifer S Knaack, et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.