Analysis of non-conjugated steroids in water using paper spray mass spectrometry

A novel strategy for the direct analysis of non-conjugated steroids in water using paper spray mass spectrometry (PS-MS) has been developed. PS-MS was used in the identification and quantification of non-conjugated (free) steroids in fish tank water samples. Data shown herein indicates that individual amounts of free steroids can be detected in aqua as low as; 0.17 ng/µL, 0.039 ng/µL, 0.43 ng/µL, 0.0076 ng/µL for aldosterone, corticosterone, cortisol, and β-estrone, respectively, and with an average relative standard deviation of ca. < 10% in the positive ion mode using PS-MS/MS. Direct detection of free steroids in a raw water mixture, from aquaculture, without prior sample preparation is demonstrated. The presence of free steroids released in fish water samples was confirmed via tandem mass spectrometry using collision-induced dissociation. This approach shows promise for rapid and direct water quality monitoring to provide a holistic assessment of non-conjugated steroids in aqua.

Steroids are a class of biologically active compounds with 17 carbon atoms arranged in a four-member ring system that are vital in a wide range of biological functions (i.e. cellular signaling, growth, development etc.) 1 . The main steroids of biological importance include cholesterol, sterols, bile acids, corticosteroids (e.g., cortisol, corticosterone) and sex steroids (e.g., estrogens, androgens) 2 . There is a growing interest in the development of new analytical tools for steroid analysis due to the potential of steroids as biomarkers for a variety of diseases, their potential misuse to enhance performance in sports [3][4][5] , and their use in animal production and to monitor animal welfare (e.g., cortisol, the primary stress hormone) 6 .
Aquaculture for food production has been valued globally at around $160 billion 7 . High stress (in fish) is reported during transport, capture and stocking procedures, which can be correlated with plasma cortisol levels [8][9][10] . This leads to increased energy consumption, decreased feeding, increased susceptibility to disease and/ or impaired survival 11,12 . Due to the increasing importance of aquaculture as a means of food production and the impact of stress on production efficiency, fish welfare is becoming an increasingly important issue for the industry. Therefore, there is a pressing need for new analytical innovations to monitor fish health and welfare. In this study we focus on establishing direct mass spectrometric approaches for rapid steroid measurements, including cortisol, which has implications for measuring stress in fish in aqua as an early warning of welfare issues 13 , as well as for environmental monitoring 14 , without causing undue stress or destruction of the animals.
The analysis and quantification of steroids is important due to their vital biological roles (i.e. development, reproduction regulations) in animals 15 . For example, β-estrone, a sex hormone, possesses potent antioxidant properties and plays important roles in maintaining normal reproductive and non-reproductive functions 15 . However, even in trace amounts, β-estrone may have adverse effects on humans and the aquatic ecosystem 16 . It is therefore essential to reliably determine trace amounts (at environmentally relevant concentrations) of β-estrone and other hormones in water. In relation to monitoring fish, in aqua steroid measurements are noninvasive (unlike blood tests) and therefore do not require any physical intervention (e.g., anesthetic injection, handling, bleeding). This ensures that results are not influenced by any stress incurred during sampling, whilst also allowing concurrent monitoring of the animal behavior 17 .

Results and discussion
First, to investigate the potential of PS-MS for steroid analysis and to establish an analytical procedure, the standard model compounds were analyzed and characterized using tandem mass spectrometry followed by quantitative analytical performance measurements. After establishing the analytical approach, 13 raw aquaculture samples were analyzed quantitatively for the four free steroids in question (i.e. cortisol (Mw 362), corticosterone (Mw 346), aldosterone (Mw 360), and β-estrone (Mw 270)).
Direct analysis and solvent optimization of free steroids in the positive ion mode using pS-MS. Figure    www.nature.com/scientificreports/ When the standard spray solvent, MeOH/H 2 O, 1:1 v/v, was doped with a weak acid (500 mM of formic acid), an intense protonated molecular ion, [M + H] + , at m/z 363 for cortisol and m/z 271 for β-estrone was obtained with the corresponding sodiated peaks diminishing ( Fig. 1A (ii), B (ii), respectively). When doping the spray solvent with a stronger acid, 0.05 mM of hydrochloric acid (HCl), the intensity of the protonated peak intensifies and any sodiation is difficult to distinguish from the noise for both cortisol and β-estrone ( Fig. 1A (iii), B (iii), respectively).
These results are in good agreement with ESI 56 and MALDI 57 . The effect of adding a dopant such as formic acid (500 mM) or HCl (0.05 mM) to the PS solvent (MeOH/H 2 O, 1:1 v/v) shows that the mechanism of ionization immediately favors the formation of the protonated molecule, [M + H] + , by proton transfer. The most likely point of protonation on the steroid molecule is at the hydroxyl (OH) group 58 . Many biologically important molecules, such as the steroids studied herein, cannot be easily ionized using the typical paper spray solvent (methanol/ water (1:1, v/v)) due to ion suppression 59 . This is because many biological samples contain salts, and many ions compete for charge 60 . The addition of a dopant such as HCl favors the production of the protonated molecular, [M + H] + , ion with little or no fragmentation (Fig. 1A (iii), B (iii)).
Tandem mass spectrometry (MS/MS) with collision induced dissociation (CID) was used to confirm the identity and the chemical structure of the analyzed steroid compounds; i.e., the protonated molecular ion, [M + H] + , of the model compounds studied using PS-MS. Figure 1A (iv), B (iv) show the CID mass spectra of cortisol and β-estrone obtained for the dissociation of the protonated molecular ion at m/z 363 and m/z 271, respectively (whereby the spray solvent was doped with HCl). Notice that, upon CID activation ( Fig. 1A  Most polar steroids can be readily protonated or deprotonated using PS-MS (in positive or negative ion mode). However, the degree of their protonation is also dependent on their proton affinity (with high proton affinity favoring formation of the protonated molecule) 61 . The addition of dopants can improve the analytical specificity and detection limits for certain analytes in complex mixtures 51 . As observed in Fig. 1, the steroids analyzed, cortisol and β-estrone, using PS-MS with an acid dopant (i.e. either formic acid or HCl), yielded dominant protonated molecular ion species, [M + H] + , with diminished sodium adducts, [M + Na] + , compared to the standard spray solvent MeOH/H 2 O (1:1, v/v). The proton affinities of the steroids studied can be estimated to be in the range 820-870 kJ mol −1 based on the method of Prof. Graham Cooks (by using group equivalent effects on the known values of simple analogues) 62 . As such, steroids can be protonated in the presence of proton donating additives such as water, formic acid or HCl ( Fig. 1) when analyzed using PS-MS. As demonstrated in this study, the ability to improve the ionization of steroids through protonation ([M + H] + ) by adjusting the dopant composition is effective. This can be particularly useful when molecular weight confirmation is required in the presence of other competing analyte(s).
By adjusting the solvent composition, the type of adduct ions obtained can be effectively manipulated in both the positive and negative ion modes. As seen in Fig. 1, this can have a significant bearing on ion formation. The various physical and chemical properties of a solvent (i.e. polarity, volatility, surface tension, etc.) play a significant role in the spray process of PS 63 and many similarities are observed when compared with conventional electrospray ionization 5 . Typically, the spray solvent composition is adjusted, as well as the sample solvent (where possible), to optimize the spray conditions for the analysis at hand 64,65 . As shown in Fig. 2, the addition of a dopant to the paper spray solvent greatly improves the signal intensity of the target analyte(s) by promoting a favorable ionization mechanism (i.e. protonation [M + H] + ).
Quantitative analysis. The free steroid model standards in 5 µL of water dried onto a paper surface and spiked with 5 µg/mL of cortisol-1α,2α-d 2 were analyzed with MeOH/H2O (1:1, v/v) doped with 0.05 mM HCl. The analytical performance of PS-MS was carried out in the MS/MS mode using CID transitions (m/z 361/315, m/z 347/293, m/z 363/327, m/z 271/158; for aldosterone, corticosterone, cortisol, and β-estrone, respectively and m/z 365/269 for cortisol-1α,2α-d 2 ). The intensity ratios of the two transitions were plotted against the known concentrations of the standard to build the calibration curves for the model compounds with a minimum of 5 points over a linear dynamic range of concentrations (0.1-10 µg/mL) covering the expected sample concentration ( Figure S3, supplementary information). The limit of detection (LOD) was determined as the concentration that produces a signal more than three times greater than the standard deviation plus the mean value of the blank, in the MS/MS mode using CID. The limit of quantitation (LoQ) was determined as the concentration that results in a signal 10 times greater than the standard deviation plus the mean value of the blank. The LODs of the four model compounds were determined to be 0.17 ng/µL, 0.039 ng/µL, 0.43 ng/µL, 0.0076 ng/µL for aldosterone, corticosterone, cortisol, and β-estrone respectively, with a relative standard deviation of ca. < 10%. While the LoQ of these steroids in aqua was determined to be 0.39 ng/µL, 0.064 ng/µL, 0.48 ng/µL, and 0.41 ng/µL for aldosterone, corticosterone, cortisol, and β-estrone, respectively, in the positive ion mode with HCl (0.05 mM) doped in the standard spray solvent, MeOH/H 2 O (1:1, v/v) ( Table 1). These results are in good agreement with previous studies, where chloride ions have been utilized in the selective ionization of carbohydrates using conventional ESI 66 (Fig. 3A-D). Gas phase ion/molecule reactions can readily take place during the ionization process during paper spray analysis. It was found that the free steroids formed chloride adducts [M + Cl] − in the negative ion mode when a diluted hydrochloric acid solution (0.05 mM) was doped in the paper spray solvent. The reactive paper spray experiment (Fig. 3) can, therefore, enhance the specificity of free steroid detection from complex sample mixtures. Moreover, the stability of the chloride adducts [M + Cl] − affords improved sensitivity, perhaps due to the higher affinity of chloride ions compared to protons. This is also supported by   www.nature.com/scientificreports/ from a high-resolution instrument 69 . The results obtained for the analysis of steroids in the negative ion mode are in agreement with those previously reported for the analysis of free and total cortisol in complex mixtures using LC-ESI with tandem mass spectrometry 66 .  Figure 4A shows the mass spectra obtained from the analysis of the steroid artificial mixture using PS-MS in positive mode. In negative mode, the ability to form [M + Cl] − adducts ( Fig. 4B) with all the studied model compounds greatly simplifies the resulting mass spectra for mixture analysis without any prior separation. This approach can potentially be very useful for field applications, when coupled with miniaturized mass spectrometers.

pS-MS analysis
Quantification of free steroids in fish water samples. The analysis of cortisol, aldosterone, corticosterone and β-estrone in real aquaculture water samples was investigated using PS-MS without any sample preparation. Thirteen (13) raw water samples were collected from fish tanks representing various conditions (not known to the authors), supplied by a fish farming company (who cannot be disclosed for commercial reasons), without any prior sample workup. A 5 μL volume from each raw sample was deposited onto individual paper substrates and analyzed using a commercial benchtop mass spectrometer in positive ion mode. Figure 5 shows a representative mass spectrum from one of the raw fish water samples (sample 8). When using the standard solvent system, a moderately intense ion at m/z 361 is observed, suspected to be a protonated aldosterone molecular ion, as well as an intense peak at m/z 383 suspected to be its corresponding sodium adduct, [M + Na] + (Fig. 5A).
In support of our suspicions as to the identity of these peaks, we then doped the spray solvent with 0.05 mM of HCl and a quite distinct mass spectrum was observed (Fig. 5B). The addition of a strong acid suppresses the sodiation ionization process favoring protonation of the steroid compounds in the mixture. Hence in Fig. 5B, only the protonated molecular ion species, [M + H] + , of aldosterone, at m/z 361, is distinguishable in the full MS mode; its identity was confirmed using MS/MS CID (Fig. 5D). The other steroid compounds (corticosterone,  Table 2 summarizes the results for the quantification of the free steroids in the raw fish water samples analyzed using PS-MS/MS. The mass spectra for samples 1-13 (except sample 8) are shown in the supplementary information in Figures S4-S15. The analysis of free steroids in fish water using MeOH/H2O (1:1, v/v) as the PS solvent does not yield suitable results due to ion suppression. This is due to the salt concentration variations, which are likely to be encountered in most biological and environmental samples. Doping the standard PS solvent, MeOH/H2O (1:1, v/v), with a strong acid such as HCl enables protonation of free steroids in complex water samples to generate stable [M + H] + species with improved sensitivity for target molecules present in complex mixtures. Although the inclusion of a fractional amount of a weak acid is common practice for MS techniques to aid protonation, the addition of a www.nature.com/scientificreports/ stronger acid, HCl, to the PS solvent for the suppression of sodiated adduct, [M + Na] + , formation was more effective. We attribute the high sensitivity of the protonated ion type to the occurrence of only one major fragment ion. In this study, free steroid hormones (aldosterone, corticosterone, cortisol and β-estrone) in water were detected using PS-MS. The detected free steroid hormones can be released into the water via three main routes; the bile, the urine and the gills 70 . We detected the presence of free steroid hormones in most of the raw samples tested. A potential factor contributing to the higher hormone concentrations detected in the samples relates to the fact that these compounds accumulate in water systems, particularly when there are very large populations (i.e., high stock density) as is to be expected in commercial aquaculture.
The method provides a good basis for relative measure of free steroids in a non-invasive manner to monitor a fish stock holistically. For example, cortisol levels were found to be relatively high in some samples. It is well known that fish which encounter a stressful stimuli launch an endocrine stress response to release glucocorticoids (i.e. cortisol) into the water through the gills, urine, feces or mucus 8 . The cortisol levels detected in the studied samples (Table 2) show some agreement with those reported in previous studies using HPLC-MS [71][72][73] . Typical stressors can occur when water renewal is limited and stocking densities are high as well as during transportation and vaccination 73 . Whilst the reported LODs herein are higher than those from other analytical methods (e.g., LC/MS, RIA), the detection of steroids at the endogenous level in water has been achieved. The detection of hormones in water using a non-invasive method (i.e., without disturbing fish) such as PS-MS at higher concentrations can open up the possibility of on-site monitoring but in order to monitor the stock holistically, it would also require information on biomass, water flow rate and tank volume to calculate actual hormone release rates 74 . The initial study presented herein demonstrates the possibility of this approach for rapid in aqua analysis. However, further optimization and validation studies would be required to extend this method for similar applications where other factors may be prominent (e.g., environmental factors, water chemistry (turbidity, pH), etc.).

conclusion
The ability to detect and characterize free steroids in raw water samples collected from fish water tanks has been demonstrated using PS-MS/MS in the positive mode. Negative ion mode operation also shows promise, however the stability of the chloride adduct prevented collisionally activated dissociation with the instrument used. Further studies could address this limitation and/or investigate the possibility of steroid analysis using paper spray in the negative ion mode using exact mass measurement with a high-resolution mass spectrometer. In positive mode, the detection limits obtained suggest that this approach is suitable for direct and rapid detection of free steroids in raw water, although, as might be expected, it registers higher detection limits than those that can be obtained from GC-and LC-MS analytical methods 24,75 . With the capability of PS-MS to perform direct analyses on unmodified samples in a simple manner, the method demonstrated in this study shows great promise for non-invasive in aqua monitoring of fish stocks holistically in aquaculture, with minimal disruption.
The method provides a basis for the relative measure of free steroids in a non-invasive manner without any intervention other than collection of a water sample. The manual approach used herein allows a result to be obtained within ~ 1 min from sampling. A key advantage is being able to monitor the holistic profile of steroid hormones secreted in to a body of water. However, a potential drawback of the free steroids measurement in water is that it effectively integrates the free steroid release of all members of a population. As such, larger interindividual differences in, for example, plasma concentrations of individuals may make a greater contribution to the overall free steroid concentration as determined from the collective. Nevertheless, the simplicity of paper spray ionization and the ability to analyze raw fish water samples without sample preparation further enhances the potential for coupling to a portable or miniaturized mass spectrometer for on-site analysis. Such a system in operation could be of great benefit for the aquaculture industry (e.g., for routine health monitoring and quality control). (HPLC grade, ≥ 95%)) were purchased from Sigma-Aldrich (UK). Thirteen (13) raw water samples were received from a commercial aquaculture company and used without preparation.
Sample preparation. Pure steroid model compounds (i.e. cortisol, corticosterone, and aldosterone, β-estrone, cortisol-1α,2α-d 2 (HPLC (grade, ≥ 95%)) were diluted to a target concentration in methanol (HPLC grade). The raw aquaculture water samples were analyzed as supplied without any sample preparation or purification using PS-MS. 5 µL of the sample was pipetted onto a cut paper triangle and analyzed directly.
pS-MS instrumentation. All steroid analysis was performed using a Thermo Scientific Velos Pro LTQ linear ion trap mass spectrometer (San Jose, CA, U.S.A.) tuned to the precursor ion of interest. The capillary temperature, capillary voltage, tube lens voltage were set to; 150 °C, 15 V, and 60% Slens voltage, respectively, in both positive and negative ion mode. Whatman grade 1 chromatography filter paper was used as the PS substrate. A DC voltage of ± 5.0 kV was applied on the PS substrate in all experiments. Figure 6 shows a sketch of the PS-MS experimental setup used in the analysis of free steroids in water. A 5 µL aliquot of each sample was deposited on a filter paper surface followed by 20 µL of 1:1 MeOH/H 2 O (v/v) PS solvent, doped with either 500 mM formic acid or 0.05 mM HCl and analyzed directly without any sample preparation. Tandem mass spectrometry (MS/MS) was used for the structural elucidation, analyte identification and quantification. MS/MS was performed on the molecular ions of interest by collision-induced dissociation (CID) using an isolation window of 0.1-1.5 Th (mass/charge units) and normalized collision energy of 30-50 (manufacturers unit). Cortisol-D2 internal standard was spiked to a total of 5 ng/µL into each test sample and calibrator prior to spotting onto the paper surface. All quantification was performed by comparing the average intensity of the analyte fragment divided by the average intensity of the cortisol-D2 fragment 365 → 269 (IS). The ion transitions used for quantification were 347 → 293, 363 → 327, 361 → 315, and 271 → 158 for corticosterone, hydrocortisone, aldosterone, and estrone, respectively.