Introduction

Pheromones can be defined as substances, or specific mixtures of substances, that are released by individual organisms into the environment, where they evoke specific and adaptive biological responses in conspecifics, and the expression of which does not require active learning (Sorensen et al., 1998). The mixture of chemicals comprising fish pheromones can be complex, and in most cases the structure and function of these compounds is still unknown.

The use of pheromones (or chemical signatures) as a management tool for conservation of species or population control is rapidly gaining popularity. However, for this approach to be successful, it is necessary to first understand the chemical structures involved in pheromone mixtures, and second, to be able to measure accurately often extremely low concentrations (ca. 10−12 M) of these chemicals in natural water samples.

One species for which the chemical signature has been intensively studied is the sea lamprey (Petromyzon marinus), where the migratory pheromone is comprised primarily of petromyzonol sulfate (PS), petromyzonamine disulfate (PADS), and petromyzosterol disulfate (PSDS) (Sorensen et al., 2005). To date, quantitation in natural water samples has been limited to PS (Fine and Sorensen, 2005).

While sea lamprey in the northern hemisphere pose a huge challenge to fisheries, especially in the Great Lakes region (Krueger and Marsden, 2007), much less is known about the ecology of the Southern pouched lamprey, Geotria australis, which is endemic to New Zealand. Once an important traditional fishery, anecdotal evidence suggests that populations have declined steadily over time (James, 2008). Recent research has focused on enhancing knowledge of the distribution and abundance of G. australis populations within New Zealand, and developing methods for monitoring populations via pheromones released by stream resident larvae (Baker et al., 2009).

Within the laboratory, the larvae of G. australis produce large quantities of PS, with markedly lower amounts of PADS, and with PSDS undetected (Stewart and Baker, in press). Thus, population monitoring of G. australis is more likely to be successful by concentrating on detection of PS in river water. However, as a generally applicable method, it would be desirable to include PADS and PSDS in the methodology.

For monitoring of fish populations (using pheromone detection), a robust, rapid, and accurate measurement of these chemicals is essential. Liquid chromatography coupled to triple quadrupole mass spectrometry (LC/MS/MS) is ideally suited to this application, as it is highly sensitive and selective in the analysis of environmental concentrations of organic water soluble chemicals. As such, we explored the suitability of a pre-concentration step followed by LC/MS/MS analysis in quantifying PS, PADS, and PSDS in a river water matrix. Here, we describe the method process and validation.

Methods

Standards

PS was purchased from Toronto Research Chemicals (P293526), while PADS and PSDS were kindly supplied by Professor Peter Sorensen, from the University of Minnesota.

Pre-concentration

Lamprey pheromones were concentrated from the water matrix by Oasis HLB solid phase extraction (SPE) cartridges (Waters Corporation). SPE cartridges (60 mg) were conditioned with methanol (2 ml) and H2O (2 ml), prior to sampling. Water samples (20 ml) were eluted through the SPE, using positive pressure. The water sample was eluted fully through the SPE cartridge and all eluent discarded. The cartridge then was removed of excess H2O by several passes of air through with positive pressure. The material retained by the SPE cartridge was immediately eluted into a 2 ml amber glass vial with methanol (2 ml). All methanol was removed under a stream of N2 gas. The sample was re-dissolved in 1:1 methanol:H2O (100 μl), sonicated briefly, and transferred to a low volume 150 μl glass insert, inside the 2 ml amber glass vial. The sample was sealed with a lid and septum for LC/MS/MS analysis. All analyses were carried out in triplicate.

Spike Recovery

PS, PADS, and PSDS were spiked into blank stream water at concentrations equivalent to 0.01 μg/l and 0.001 μg/l. Water from a stream with no known lamprey larvae (Donny Park Stream, a tributary of the Waikato River) was used for spiking and procedural blanks. Samples of stream water (100 ml) were spiked with each standard (100 μl at 1000-fold desired concentration), to keep methanol concentration negligible (0.1% v/v), and avoid compromising the performance of SPE.

Procedural Blanks

Procedural blanks were used for stream water samples and were processed as above.

Post SPE Spikes

To assess matrix interference procedural blank samples were spiked with standards (9.1 ng/ml) and analyzed by LC/MS/MS.

LC/MS/MS Analysis

Quantitation of PS, PADS, and PSDS was achieved on an Agilent 6460 Triple Quadrupole mass spectrometer with an electrospray ionization (ESI) source coupled to an Agilent binary pump HPLC system (1200 Series Binary Pump SL, HiP-ALS SL+ autosampler and 1200 series degasser). Samples were analysed on an Agilent Zorbax 1.8 μm, 2.1 × 50 mm SB-C18LC column, using a gradient of mobile phase A (0.01 M ammonium acetate in Type 1 water) and mobile phase B (acetonitrile). Flow was 0.3 ml/min with an initial proportion of 20% B holding for 0.5 min, a linear gradient to 30% over 0.5 min, 60% over 2.5 min, 100% over 0.5 min, holding for 0.5 min, and back to 20% B over 0.5 min and holding for 1 min, a column temperature of 50°C was maintained. PS and PADS were analysed in positive ion mode, while PSDS was analysed in negative ion mode. PS was observed as the ammonium adduct at m/z 492.4 ([M+H+NH4]+) with specific transitions to m/z 359.3 for the quantifier ion (Quant) and m/z 412.4 for the qualifier ion (Qual). PADS was observed as the trihydrogen adduct ([M+3H]+) at m/z 705.4, with the Quant ion at m/z 509.3 and the Qual ion at m/z 607.4. PSDS was observed as the doubly charged anion ([M]2−) at m/z 294.3, with the Quant ion at m/z 491.3 and the Qual ion at m/z 97 ([SO4H]−).

Results

Quantification of Pheromones

Spike recoveries for PS were 79.2 (2.5)% and 76.4 (8.1)% [(% Relative Standard Deviation (RSD)] for concentrations of 0.01 μg/l and 0.001 μg/l, respectively. Spike recoveries for PADS were 67.1 (5.5)% and 65.5 (3.0)% (% RSD) for concentrations of 0.01 μg/l and 0.001 μg/l, respectively. Due to poor fragmentation of PSDS in spike recoveries, it could only be quantified for the 0.01 μg/l spike experiments, with an average recovery of 141 (10)% (% RSD).

Post-spiking of procedural blanks showed a 92–95% response of the signal expected for all three pheromones, suggesting that a typical river matrix did not result in significant suppression or enhancement when using this method, at the higher 0.01 μg/l spike concentration.

The triple quadrupole analysis afforded instrument practical quantitation limits (PQL) of 0.05 μg/l for PS and PADS and 0.5 μg/l for PSDS.

Discussion

Previous methods for quantitation of lamprey migratory pheromones have been restricted to PS and allocholic acid (ACA) by using methods based around induced fluorescence by enzymatic oxidation of the 3α-hydroxy functional group (Polkinghorne et al., 2001), or direct infusion mass spectrometry by using selected ion monitoring (SIM) (Fine and Sorensen, 2005). The method of Polkinghorne et al. (2001) is time consuming (160 min), reasonably insensitive (detection threshold >0.01 μg in a clean background), requires a custom made column, is not selective to specific chemicals of interest (will detect any 3α-hydroxy bile acid), and will not detect PADS or PSDS. The method of Fine and Sorensen (2005) uses an ion trap mass spectrometer, which is incapable of fragmenting PS and so relies on a narrow SIM window to detect and quantify this compound. Furthermore, with no chromatographic fractionation, matrix effects can be large and there is no retention time confirmation.

The LC/MS/MS method presented here has many benefits over existing methods for quantifying lamprey migratory pheromones. It uses a short highly efficient HPLC column to provide retention time confirmation of the compound of interest, and as such is extremely rapid, with a total run time of 7 min. It uses defined concurrent fragmentation of each compound of interest to form ions for quantitation (Quant) and ions to qualify (Qual) the existence of that compound (see Fig. 1). The method is further validated by analysis of Quant/Qual ion ratios and is extremely sensitive, with instrumental practical quantitation limits (PQL) of 0.05 mg/l for PS and PADS and 0.5 mg/l for PSDS. The PQL is defined as the minimum concentration of which quantification can be achieved with known level of accuracy and precision. A pre-concentration factor of 200 equates to “in water” PQL of 0.25 ng/l for PS and PADS and 2.5 ng/l for PSDS, corresponding to molar concentrations of 5 × 10−13 M, 7.5 × 10−13 M and 6.25 × 10−12 M for PS, PADS and PSDS, respectively. However, greater pre-concentration steps also are applicable, and a 4000-fold pre-concentration—as has been used in water sampling of streams (Stewart and Baker, 2011)—equates to “in water” molar PQLs of 2.5 × 10−14 M, 3.75 × 10−14 M, and 3.125 × 10−13 M for PS, PADS, and PSDS, respectively. The analysis is fully automated, is not labor intensive, and can process many samples in a short period of time. Furthermore, the suite of chemicals of interest can be expanded manifold without significantly increasing the method run time or impacting on other chemicals currently being analysed.

Fig. 1
figure 1

LC/MS/MS chromatograms of PS, PADS and PSDS. a PS standard at 2 ng/ml (top trace) and spikes at 10 ng/l and 1 ng/l (bottom trace); b PADS standard at 2 ng/ml (top trace) and spikes at 10 ng/l and 1 ng/l (bottom trace); c PSDS standard at 2 ng/ml (top trace) and spike at 10 ng/l (bottom trace): Note different time scales for PSDS

Full size image

The analysis of PSDS by this method was not as robust as that of PS and PADS, with observed sensitivity reduced by 90%. This was largely a function of the poor fragmentation yield from the parent ion (m/z 294.3) to the quantifier ion (m/z 491.3) and low specificity of the relatively low molecular weight qualifier ion (m/z 97), leading to significant background noise from the matrix (Fig. 1c). Further avenues could be investigated to improve the detection limit of PSDS. However, this is the first described method for quantitation of PSDS and with suitable pre-concentration (4000×), sub picomolar quantitation can still be achieved.

A similar method for quantitation of the male sea lamprey sex pheromone, 3-keto-PS has been described recently (Xi et al., 2011). Like the current method, the method of Xi et al. uses triple quadrupole mass spectrometry for analysis. However, the authors noted that a mixed mode ion exchange cartridge was necessary to minimize matrix interference. Fragmentation of 3-keto-PS was not selective, with a daughter ion observed at m/z 97, as was observed for PSDS in the current method. Such a non-selective ion fragment likely accounts for the high matrix interference observed in both our method for PSDS and that of Xi et al. The method of Xi et al. uses negative ion mass spectrometry to quantify 3-keto-PS. We found that positive ion mass spectrometry—facilitated by ammonium acetate in the mobile phase—gave more useful fragmentation ions for PS and PADS than negative ion, leading to reduced matrix interference and associated lower PQL.

The methodology described in this paper is suitable for quantifying environmental concentrations of migratory lamprey pheromones. In addition, the methodology can be modified easily to measure other organic chemicals present in natural waters, such as hormones, pheromones, and metabolic by-products excreted by fish, provided suitable standards are available to achieve accurate quantitation. This method (or related methodologies) hopefully will provide a useful tool in species monitoring and management, and we aim to use this technique to develop an assay for assessing lamprey populations within New Zealand streams.