Identification of Pharmaceuticals in The Aquatic Environment Using HPLC-ESI-Q-TOF-MS and Elimination of Erythromycin Through Photo-Induced Degradation

As result of the solid phase extraction, a yellowish to dark green solution was obtained in all cases, which indicated the presence of chlorophyll-containing substances (Figure 1). Pharmaceuticals contained in this water sample would not lead to visible coloration since their concentration and their absorbance would generally be too low. Instead, the occurrence of pharmaceuticals needs to be analyzed using HPLC and high-resolution mass spectrometry.

In non-targeted analysis, a HPLC-ESI-Q-TOF-MS was used because of its outstanding mass accuracy allowing to obtain the accurate mass for each compound ion. The mass-detected chromatogram of the performed analysis was represented as a base peak chromatogram (BPC), which displays the most intense peak of each mass spectrum recorded in the course the chromatographic separation. The example shown in Figure 2 presents the BPC of a water sample from the river Rhine.

The BPC contained more than twenty-five peaks reflecting different m/z values, hence different compounds, seven of which were marked in the BPC. Since the substances were unknown a priori, the first step to their identification usually consists of deriving the molecular formula. This is accomplished through accurate mass and isotopic pattern provided by the TOF detection, although the isotopic pattern may not be observed in all cases due to low sample concentrations in environmental samples. With the help of public database, such as Pharmaceuticals in the Environment by the German Environment Agency (UBA) containing approximately 630 compounds, a preliminary identification of a small group of candidates is often successful. For a final proof, either comparison to commercially available reference standards may be performed or MS/MS fragmentation patterns may be considered (Figure 3).

In this work, comparison to standards with respect to retention time accounted for the identification of pharmaceuticals very often found in German surface waters. These substances include metoprolol, a β-blocker, carbamazepine, an analgesic, and the macrolide antibiotic erythromycin A and its derivate anhydroerythromycin A. Erythromycin serves as example further investigated in this study. The studied Rhine river sample had a pH of 7.6 and an average temperature of 16.5 °C. At this pH, anhydroerythromycin would also be expected to be present in the water sample. For the detailed analysis, the extracted ion chromatograms (EICs) of the water sample were compared with the reference standards (Figure 4).

The comparison shows good agreement between the retention time for metoprolol, carbamazepine and anhydroerythromycin and the observed analytes. The EIC of the reference standard anhydroerythromycin displayed two peaks, hence two compounds where dehydration had occurred at two distinct sites of erythromycin. Yet, only one anhydroerythromycin isomer was identified in the Rhine river sample. Erythromycin itself was only present in traces. Therefore, no MS/MS spectrum could be obtained. The accurate masses for the antibiotic and its dehydrate are given in Table 2. Using EIC, thus m/z value and retention time, metoprolol, carbamazepin, erythromycin and anhydroerythromycin could be identified in the Rhine river sample.

With respect to the aquatic environment, it is important to prevent pharmaceuticals from passing through wastewater treatment plants and entering surface waters. In the quest for an efficient elimination, UV-C irradiation experiments at different pH values were carried out for erythromycin as example. Concentration-time (c-t) diagrams were recorded using mass-area vs. time plots derived from EICs. The degradation was described according to equation 2. Erythromycin consists of erythromycin A and B and anhydroerythromycin A, with two isomers of the latter. The c-t curves of erythromycin A and their computational fits are shown in Figure 5. At pH 7,accelerated degradation was observed. This applies to all four compounds studied, data not shown. As a consequence, photo-induced degradation of erythromycin should be carried out around neutral pH. In the case of the Rhine river sample, pH adjustment was not required.

Photodegradates of the pharmaceuticals were also identified at all three pH values. An overview of these photodegradates with their corresponding structure proposals is given in Table 3. For the kinetic analysis of the photodegradates, the product with m/z = 720 serves as an example. Photodegradates can often be described as reaction intermediates. Therefore, the photodegradates were described in terms of aconsecutive and subsequent follow-up reaction. The decision between the resulting types of intermediates is based on the goodness of the fit computed with suitable software, where the coefficient of determination (R²) and the residual mean-squared error (RMSE) were taken as the criteria. Due to the fact that erythromycin is acid-instable, degradates as would occur upon irradiation were present prior to irradiation. The resulting effect on equations 3 and 4 was a finite starting concentration. Hence, a factor was added to the equations. Figure 6 shows experimental data and fits computed according to equation 3 and 4.

This example of an intermediate demonstrated the concentration increase with a sigmoidal rise followed by an exponential decay. This is indicative for a subsequent follow-up reaction intermediate. A consecutive reaction intermediate does not show the sigmoidal increase. Statistical quality parameters also indicated the slightly superior agreement of fit according to the subsequent follow-up reaction model. The coefficient of determination R² of the consecutive reaction was 0.9898 and thus lower than that of the subsequent follow-up reaction being 0.9976. Therefore, the examined photoproduct was interpreted as intermediate of a subsequent follow-up reaction. The k-values resulted from the computational fit as well, the half-life was calculated following equation 5. All relevant kinetic parameters are collected in Table 3.

The fastest degradation was observed at pH 7, followed by pH 9, while the slowest degradation was found for pH 3 (Figure 5). This finding also applied to the formation and degradation of the photoproducts. Three photodegradates were observed. Their m/z values were 750.46 corresponding to Ery F, 720.45 to Ery C and 192.12 to DPEry192, a glycosidically bound sugar of the erythromycin structure (Figure 7). No degradation of the photoproduct could be observed for DPEry192 at pH 3 and 9 and for Ery F at pH 9. In these cases, the irradiation time was not sufficient long to observe total degradation of the intermediate product. Nevertheless, the formation rate constant could be determined by using equation 5, which corresponds to a final product.

Figure 1. Comparison of the samples from the Rhine river after SPE (left) and ultrapure water(right) treatment. The green coloration is indicative for chlorophyll-containing substances. Please click here to view a larger version of this figure.

Figure 2. BPC of a water sample after SPE measured with HPLC-ESI-Q-TOF-MS. All chromatograms were normalized to the highest peak. Illustrative m/z-values as obtained from the corresponding MS spectrum are marked. Please click here to view a larger version of this figure.

Figure 3. Q-TOF-MS spectrum of erythromycin A (bottom) and the MS/MS spectrum of the ion m/z = 734.4689 (top). The spectra show the quasi-molecular ion of erythromycin A with its isotopic pattern and the fragments at an applied collision energy of 30 eV. Please click here to view a larger version of this figure.

Figure 4. Normalized EICs of (A) metoprolol, (B) carbamazepine, (C) erythromycin A and (D) anhydroerythromycin A in a Rhine river sample (blue) and in ultrapure water from reference compounds (red). The retention times of the reference compounds and the ones of the pharmaceuticals in the water sample are the same. The signal-to-noise ratios of metoprolol (A) and anhydroerythromycin (D) are higher than those of carbamazepine (B) and erythromycin (C), which indicates the latter were present only in traces. Please click here to view a larger version of this figure.

Figure 5. Normalized concentration-time curves of the photodegradation of erythromycin A at pH 3 (red), pH 7 (green) and pH 9 (blue). Solutions were irradiated for 10 min. At pH 7, erythromycin was completely removed from the sample. The concentration-time curves could be described using first-order kinetic equations. The kinetic rate constants were 0.10 (pH 3), 0.59 (pH 7) and =0.21 (pH 9). Please click here to view a larger version of this figure.

Figure 6. Comparison of the fits of the concentration-time curves of the photoprodegradates of erythromycin with m/z = 720 at pH 9 following equations 3 (A) and 4 (B). Goodness of the fit of the consecutive reaction (A): R² = 0.9898, RMSE = 4.645E+04, and of the subsequent follow-up reaction (B): R² = 09976, RMSE = 2.366E+04. Please click here to view a larger version of this figure.

Figure 7. Structure of erythromycin A, erythromycin B and anhydroerythromycin and their photdegradation products. This figure has been modified from Voigt et al.²⁷. The products were formed after 10 min of UVC-irradiation and identified using HPLC-Q-TOF-MS and MS/MS. Please click here to view a larger version of this figure.

Table 1. Conditions and parameters used for HPLC-ESI-Q-TOF-MS analysis of pharmaceuticals in water matrices. It is advisable to introduce a rinsing step between the chromatographic runs through running a sample of pure ultrapure water between two analyses or through extending the run time of the chromatographic method in order to elute all substances.

Table 2. Pharmaceuticals found in the Rhine river sample with their retention time, theoretical and observed [M+H]⁺ and their structure. The ESI mode was set to positive, so that [M+H]⁺-ions were detected. The retention time may vary minimally for usual experimental known reasons.

Table 3. Kinetic rate constants and corresponding half-lives of the degradation of erythromycin and its photodegradates adapted from Voigt et al.²⁷. Erythromycin consists of erythromycin A, erythromycin B and two forms of anhydroerythromycin. Three photodegradates were observed. There are referred to as Ery F, Ery C and DEry192.