«Determination of tetracycline residues in soil by pressurized liquid extraction and liquid chromatography tandem mass spectrometry Vicente Andreu · ...»
Experiments were performed to determine recoveries of the antibacterial agents CTC, OTC, TC and DC for the tandem SPE (SAX+Strata-X) clean-up step only. Sample matrix was obtained by extracting control samples, using the described PLE method. These PLE extracts were fortified with antibacterial agents at two concentration levels (1.25 and 12.5 μg l−1), corresponding to extracts obtained from extraction of soil samples with antibacterial agent contents of approximately 10 and 100 μg kg−1 soil, respectively. The fortified samples were passed through the SAX–Strata.X SPE cartridges as described above.
Day-to-day variations for the extraction procedure were determined, by repeating the recovery experiment for concentration levels of 10, 50 and 100 μg kg−1 soil after 3 days.
The linearity of the analytical methods was proved building the calibration curves for each compound, using soil extracts without TCs spiked at six concentration levels (10, 25, 50, 75, 100 and 500 μg kg−1 soil) with each TC. Each level was prepared in triplicate. The linear regression analysis was carried out by plotting the peak area ratio of analyte and IS versus the analyte concentration.
Furthermore, the linearity range for the entire soil extraction procedure was determined by fortified soil samples at the same six concentration levels that previously. Triplicate samples were made for concentration levels.
The LOD was estimated at a signal-to-noise ratio of 3, while the LOQ value was estimated by using a signal-to-noise ratio of 10. LODs and LOQs were obtained by the transition with higher signal/noise in SRM mode. For the LOQ, the confirmatory transition should be, at least, visible in the chromatogram. Once evaluated, three samples were spiked at the estimated levels and extracted according to the proposed procedure to ensure their feasibility.
Results and discussion Soil extraction method optimisation Extraction solvent for extracting the TCs from soil was selected in accordance with their physicochemical properties. Many different soil-adsorption mechanisms, such as hydrophobic interactions, hydrogen bonding, complexation and cation exchange, may affect the extraction of the compounds from soil [18,22,31]. TCs form strong complexes with di- and trivalent cations in the clay mineral inter-layers or with hydroxy-groups at the surface of the soil particles [19-21]. As a starting point, a sample pre-treatment by homogenization of soil with EDTA-Na2 washed sand as dispersing agent followed by PLE using water as extractant, was applied as previously employed for food analysis [27,32].
The influence of the sorbent used for the homogenization and dispersion of soil was investigated by performing the homogenization with alumina (neutral and basic), Florisil®, silica and sea sand, all of them washed with EDTA-Na2. The washing of the material was carried out because previous studies performed in food demonstrated that EDTA-Na2 washed materials always provide better recoveries . EDTA-Na2 deactivates metal impurities present in the surface and, probably, chelates also metals present in soil facilitating the TCs decomplexation. Silica and sea sand provided the better recoveries for all TCs (up to 40 %). Alumina, neutral or acid, does not recover these compounds and Florisil® provides good recoveries for OTC, CTC and their epimers but fails in recovering TC, e-TC and DC.
In order to obtain the highest possible concentrations of the TCs in soil extracts, the performance of the PLE-system was investigated for soil samples weighing up to 25 g. However, samples greater than 5 g were difficult to analyze owed to the larger number of matrix producing peaks. To increase contact-surface between soil particles and water and prevent metal complexation, as well as clogging of the extraction cell, 5 g EDTA-Na2 washed sea sand was mixed into 5 g soil sample before extraction. In development and optimisation of the PLE method several settings need to be considered such as pressure, temperature and number of solvent cycles. The final operating parameters for the validation of the method are listed in the Experimental section (see PLE extraction).
Chemical transformation processes of TCs, such as isomerization and epimerization, have been reported giving rise to structurally related compounds . The possible inter-conversion between TCs and their 4-epimeric (4-eTCs) forms during the extraction procedure was checked by spiking different soil samples only with the TCs or with the 4-eTCs. No inter-conversion between the TCs and its 4-epimers was observed, even through the extraction was performed at 70 ºC.
The method was compared with two methods based on PLE recently reported by Jacobsen et al.
 and O’Connor et al.  (Fig 1). The former procedure is based on mix 10 g soil sample with 10 g Ottawa sand before added to the extraction cell of 33 ml. The extraction buffer consisted of a 50:50 (v/v) mixture of methanol and 0.2 M citric acid in water with pH adjusted to 4.7 with NaOH.
Extractions were performed at room temperature to avoid that the TCs were converted to their epior anhydro form. Recoveries, achieved using this method, were above 40%, which were comparable to recoveries for the proposed combination. The latter protocol involves extraction of 5 g of soil with 50:50 (v/v) methanol-acetate buffer (pH 8). The percent recoveries of the optimized PLE method varied between the soils and ranged from 22-99%, depending on soil type, and more specifically on clay content. Comparing the three extraction methods, i.e., the results shown in Fig.
1, all of them provided acceptable recovery levels of the studied TCs with overall recoveries of 77 % for the method proposed by Jacobsen et al. , 74 % for that proposed by O’Connor et al.  and 78 % for the method developed in this study. The main advantages of the proposed procedure for the TC extraction from soils are that only uses water as extractant and the stability of the compounds through the extraction procedure.
Clean-up and pre-concentration using SPE Solid phase extraction (SPE) is traditionally used as a clean-up technique for extracts of environmental samples. The two reported procedures for tetracyclines used combination of SAX and HLB (polymeric) or SAX and StrataX SPE sorbents, respectively. These sorbents act as both clean-up and pre-concentration. The SAX column reduces matrix interferences by adsorbing anionic humic particles from the soil extracts, avoiding contamination, blocking and overloading of the HLB or StrataX sorbent. At the extraction pH, the TCs are overall neutral or cationic and are therefore not retained on the SAX cartridge, while the polymer based HLB or StrataX cartridge simultaneously retains neutral polar and non-polar compounds, including the studied TCs. A study of the results obtained using HLB and Strata X cartridges, individually and the combination SAX and HLB and SAX and StrataX was carried out. HLB and StrataX cartridges, showed, as expected, an appropriate analyte isolation and strong adsorption of matrix compounds, since TCs were quantitatively eluted with 2 ml of methanol. The extract obtained using either HLB or StrataX cartridges showed that extract still has an intense yellow-brown colour indicative of co-extraction of soil components. Additional purification of the sample by SAX resulted in a considerable decrease in the humic acid content providing almost transparent extracts. The interpretation of the chromatograms with low concentrations of the analytes improved at the same time as the lifetime of the LC column was prolonged.
Recoveries and corresponding 95% confidence intervals for the different SPE platforms, at two concentration levels (approximately 1.2 and 12.5 μg l−1 in the final extract) are shown in Fig. 2. The recoveries obtained using the SAX-StrataX systems were slightly better than those obtained with the SAX-HLB because of this, the former was selected for the definitive method.
Validation of the soil extraction procedure Figs. 3 and 4 display a typical total ion current (TIC) and mass chromatograms obtained for a soil sample spiked with the four TCs plus the IS and with the three 4-epimers and the IS, respectively.
The TIC chromatogram is a summation of the ion signal generated by all the precursor → product ions transition showed in Table 1. Chemical transformation processes of tetracyclines, such as isomerization and epimerization, have been reported giving rise to structurally related compounds.
For instance, CTC is converted to isochlortetracycline (iCTC) under alkaline conditions, while the epimerization has been found to be catalyzed in acidic solutions in a pH range from 2 to 6 . The mass chromatograms (Fig. 1B-E and Fig. 2B-D) show that there is no inter-conversion between TCs and their 4-epimeric forms during the extraction procedure. In the present study, epimerization was observed in both, standard and extract of the spiked samples, after preserving them more than one month as indicated in the experiment section. The formation of other degradation analogues was observed for CTC, which also displayed another peak that also forms epi-analogues. This peak was previously reported and tentatively identified as iCTC based on the absence of the transition 679 → 444, which indicates the lack of hydroxyl group of the CTC .
Calibration curves were obtained by spiking soil extracts without TCs and for the entire soil extraction procedure, by fortifying soil samples from control samples with TCs at six concentration levels. Validation results are shown in Table 3. Linear calibration curves were obtained for all compounds and for both soils, with regression coefficients (R2) in the range of 0.99 and linear for the complete range tested. Slope values were similar indicating only small differences attributed to the losses of TCs during the extraction procedure. Values obtained for LOD and LOQ for the TCs were in the range 1–3 and 3–10 μg kg−1, respectively, which covers the range expected for environmental soil samples and make the analytical method applicable for analysing most of them for the studied TCs.
Mean recoveries and corresponding confidence intervals for the PLE-extraction of six replicate soil samples are listed in Table 4. Recoveries for TCs were between 82 and 99 % and the relative standard deviations (RSDs) on six replicate samples were between 8 and 15 %. For the TCs (OTC and CTC) recoveries of approximately 71–78 % were achieved, which is lower than recoveries obtained for the SPE method, indicating that the compounds are not fully extracted from the soil.
This is probably due to the many sorption mechanisms involved in the binding of TC to soil, resulting in very strong sorption (Table 1). This corresponds well with previous studies indicating that TCs are sorbed to the clay fraction of the soil by complexation and hydrogen bonding and bound to acid sites in the organic fraction [17-21].
Field study Frequency, concentration, and identity of TCs found in the soil samples analyzed are outlined in Table 5. Of the 100 samples analyzed, TCs were detected in 25. The most commonly detected TC was OTC, followed by CTC, TC and DC. OTC was detected in 21 samples at levels ranging from 15.7 to 105.4 µg kg-1, CTC in 11 samples in the concentration range of 5.8-34.4 µg kg-.1, TC in 9 samples at concentrations from 18.8 to
64.3 µg kg-1 and DC in 4 at levels ranging from 12.1–45.7 µg kg-1.
On the co-occurrence of TC residues, 12 samples contained two TCs that were in seven cases OTC and CTC, in four OTC and TC and in one OTC and DC. Three samples contained 3 TCs that were OTC, CTC and TC. Only one sample contains the four TCs studied.
The isomerization of the different TCs to their epimers in soil samples was also checked. Data reported in Table 5 were calculated for the sum of both, the TC and the 4-epimer. However, Fig. 5 illustrates a chromatogram of one sample containing OTC and CTC (sample 1b of Table 5). The results shows that OTC epimerizes in low percentage, less than 25 %, comparing with CTC that shows its isomeric conversion product, iso-CTC and their 4-respective-epimers. As can be observed in the chromatogram, the conversion of CTC to iso CTC also take place in an important percentage.
TC is also epimerized in a mean value between OTC and CTC. The quantification of iso-CTC was not possible because we do not have available the iso-CTC standard. These results indicated the need for a further study on TCs metabolites and the difficulties to perform it, starting by the lack of analytical standards.
These data confirm the recent findings reported in the literature [12-15] on that TCs occur in relatively high concentrations and persist in the environment after repeated fertilizations of farmland with liquid manures or sludges. However, these data also include several novelties, such as: (i) the finding of DC; (ii) the observation of co-occurrence of TCs in soil samples that have received an unknown treatment with sludge -this co-occurrence can be by the coexistence of different TCs in the manure or sludge or by the soil treatment with different sludges, and (iii) the different isomerization patterns of TCs in real soil sample.
Conclusions Simultaneous extraction of TC, CTC, OTC and DC from soil was carried out using hot water PLE, clean-up and concentration by SPE and analysis by LC–ESI-MS/MS. Recoveries, LOD and LOQ for the extraction procedure were satisfactory, demonstrating its applicability for simultaneous determination of TCs from soil. It can be concluded that the proposed PLE-SPE method is an interesting alternative extraction technique for the determination of TC residues in soil because it provides similar results to other techniques, reduces the use of organic solvents and complex buffers and does not need pH adjustment. A pre-concentration and cleanup step is necessary because the large amount of co-extracted humic and fulvic acids. SAX-StrataX was found to be an efficient clean-up step that selectively remove humic and fulvic acids. The proposed method compares well with the results obtained by the other recently reported procedures and present the advantage of eliminating the use of organic solvent as extractants. Considering the savings in time and solvent consumption, which are both diminished by 90 %, PLE is an attractive alternative for extracting TC residues from soil.
The application of the method made the detection of significant amounts of persistent TCs in the soil possible. The study pointed out that TCs, which are frequently used worldwide, are persistent in soil in significant amounts, and that these substances represent an actual environmental problem.