A large number of pharmacologically active compounds has been detected in the aquatic environment during the last decade [1] and this gives rise to many concerns [2]. The substances have been disclosed by accident, clever search or elaborate mass spectrometry. Nowadays a broad spectrum of drug parent compounds can be screened for [3], yet the knowledge about the presence of isomers and transformation products (metabolites, conjugates, degradates) is still limited.

Non-target Analysis
A lot of information about the appearance of metabolites in urine and feces can be taken from the registration files of the pharmaceuticals. Those compounds will be diluted by wastewater, and additional transformation products of the parent and their physiological metabolites are formed (i) during wastewater treatment processes and (ii) while traveling the surface waters, e.g. by sunlight. Their detection and the discovery of other, probably novel compounds, require non-target analysis [4]. This is achieved by the employment of high-resolution mass spectrometers, which are still costly instruments. In return, these instruments deliver large data sets; in other words non-target analysis often turns into the famous search for the needle in the haystack. Obviously, alternative methods for non-target screening are welcome.

Immunochemical Methods
For many pharmaceuticals antibodies have been developed to analyze the drug parent compounds in urine and blood, e.g. in clinical studies or to produce data for the compounds' registration. Many of these antibodies are commercially available. Our working group uses these antibodies to create ELISA (enzyme-linked immunosorbent assay) immunoassays, for the determination of carbamazepine [5] for example.

Carbamazepine (CBZ) is a psychoactive drug [6], an anti-convulsant used in the treatment of epilepsy and an anti-depressant. It is rather persistent and hardly degrades in regular wastewater treatment systems [7]. Due to the high incidence of its detection in wastewater and in surface water it even has been proposed as a marker for human contamination of water bodies [8].

Our assay allows for rapid, sensitive and robust determination of CBZ in all relevant types of water.

Good correlation with a reference method (LC-MS/MS) was achieved. A detection limit of 20 ng/L for carbamazepine is reached in surface water, detection limits for wastewater are higher. All structurally conceivable and available compounds have been assayed to determine the selectivity of the antibodies during the validation of the ELISA. The ELISA shows considerable cross-reactivity (or non-selectivity) for 10,11-dihydro-CBZ, epoxy-CBZ and some to 2-hydroxy-CBZ.

LC-ELISA
We developed a low-cost approach for non-target analysis using antibodies as a tool to support the detection of hitherto unknown compounds that are structurally related to the target analytes (fig. 1, with typical experimental details). To analyze this, highly concentrated samples were prepared from heavily contaminated water sources, especially wastewater. The enrichment factor in these non-selective solid-phase extraction runs were up to 4,000-fold. In HPLC runs of such concentrated samples unselective detectors like UV/Vis, cannot be used to identify further compounds (see, e.g. fig. 2A). LC-MS/MS runs in scan mode would deliver thousands of peaks and even with HR-MS detectors it is not easy to detect "reasonable" compounds. Setting the MS parameters for the selective monitoring of ions or mass transitions (SIM, MRM modes) results in almost background-free chromatograms which allow for sensitive detection (fig. 2B,C) - but only for known compounds for which MS-parameters have been established or can be taken from databases.

We use the selectivity of antibodies instead of narrowing down the regions that are interesting, i.e. related compounds eluted from the HPLC column.

Fractions are collected every 15 - 60 sec. into a glass microtiter plate and after evaporation of the solvent and reconstitution in water or a suitable buffer the individual wells are assayed for "binding": LC-ELISA. One plate is sufficient to collect a 25 - 30 min. run. By ELISA we determine the relative analyte equivalent concentration (in %) in the fractions. Plotting them over the retention time results in nice ELISA­ grams (fig. 2D-G).

Cetirizine Uncovered
An unidentified compound was detected by the CBZ ELISA and its exact mass determined by HR-MS (UPLC-ESI/Q-TOF MS). It turned out that the compound with m/z = 389.168 is not a transformation product of carbamazepine but is cetirizine, which belongs to a series of piperazine/biphenyl compounds dissimilar to CBZ but able to bind to the antibody [9]. When superimposing the structures the norchlorcyclizine skeleton of cetirizine shows a high degree of congruence, despite of the missing bridge (fig. 3). Cetirizine binds stronger to the antibody than CBZ at neutral pH, but, since cetirizine`s pKa is at 8.0 its cross-reactivity is pH-dependent (fig. 2D-G). At pH 10.5 the binding becomes negligible and the assay shows increased selectivity resulting in improved accuracy [9]. Cetirizine is an anti-histaminic with broad use to attenuate hay fever. It seems to be a compound of wider prevalence, too, recently it was found in Portugal [10].