Complementary Mass Spectrometry and Bioassays

Trends Trends in Analytical Chemistry, Vol. 28, No.
5, 2009
Complementary mass spectrometry

and bioassays for evaluating
pharmaceutical-transformation
products in treatment of drinking
water and wastewater
Jelena Radjenović, Mira Petrović, Damià Barceló
This article summarizes existing studies on complementary mass spectrometry (MS) and bioassays for evaluating treatment
processes for drinking water and wastewater [e.g., chlorination, ozonation and advanced oxidation processes (e.g., TiO2 phot-
ocatalytic oxidation, photo-Fenton)] in eliminating pharmaceutical residues. The toxicities of transformation products of phar-
maceuticals were rarely determined individually, so we also cover research papers dealing with the total toxicity of the treated
solution. To illustrate possible strategies when performing qualitative analysis, we discuss studies based on the use of convent-
ional MS methods, and advanced MS methods involving multiple MS (MSn) experiments [i.e. triple quadrupole (QqQ-MS) and ion
trap (IT-MS)] and accurate mass measurements [i.e. time-of-flight (ToF-MS) and LTQ Orbitrap-MS].
ª 2009 Elsevier Ltd. All rights reserved.
Keywords: Advanced oxidation process; Bioassay; Chlorination; Degradation product; Drinking-water treatment; Mass spectrometry; Ozonation;
Pharmaceutical; Transformation product; Wastewater treatment
1. Introduction and diazepam) to no metabolization (e.g.,

Jelena Radjenović,
Mira Petrović,
iopromide and diatrizoate). The cyto-
Damià Barceló In the vast array of contaminants of chrome P450-mediated phase-I metabo-
Department of Environmental anthropogenic origin reaching our water lism of drugs can result in more reactive
Chemistry, IDAEA-CSIC, c/Jordi supplies, pharmaceuticals are among compounds, as a result of introducing a
Girona 18-26, 08034 those most continually input into the functional group or unmasking one (e.g.,
Barcelona, Spain
environment, especially over-the-counter hydroxylation, epoxidation, reduction and
Mira Petrović* drugs. Most modern drugs are small or- hydrolysis). Furthermore, many phase-I
Institucio Catalana de Reserca i ganic compounds with molecular weight products undergo subsequent phase-II
Estudis Avanzats (ICREA), below 500 Da and are designed to have reaction with an endogenous substrate
Barcelona, Spain specific pharmacological and physiological (e.g., glucuronic acid, sulfuric acid, acetic
Damià Barceló
effects at low doses, so they are inherently acid or an amino acid), which leads to the
Institut Català de Recerca de potent and often have unintended effects formation of highly polar conjugates that
lÕAigua (ICRA), Parc Cientı́fic i in wildlife. are excreted in the urine or bile.
Tecnológic de la Universitat de After oral, parenteral and/or topical Such conjugates can be cleaved back by
Girona, Pic de Peguera, 15, administration, a pharmaceutical is ex- the bacterial enzymes during biological
17003 Girona, Spain
creted via liver and/or kidneys as a mix- wastewater treatment, releasing the par-
ture of parent compound and metabolites ent molecules, so a large fraction of
that are usually more polar and hydro- pharmaceuticals will be discharged into
*
philic than the original drug [1]. The ex- wastewater unchanged or in the form of
Corresponding author.
Tel.: +34 934 006 100;
tent of drug metabolism in the human degradation products that are often hardly
fax: +34 932 145 904x432; body varies greatly, ranging from com- eliminable in conventional wastewater-
E-mail: [email protected] plete metabolization (e.g., carbamazepine treatment plants (WWTPs) [2]. The
562 0165-9936/$ - see front matter ª 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.trac.2009.02.006
Trends in Analytical Chemistry, Vol. 28, No. 5, 2009 Trends
extensive literature on the occurrence of pharmaceuti- In attempts to reduce the discharge of pharmaceutical
cals in WWTP effluents reinforces concerns about dis- residues into the environment, chemical-oxidation pro-
charged pharmaceutical residues that may end up in the cesses [e.g., ozonation and advanced oxidation processes
water supply, potentially resulting in adverse effects for (AOPs)] have been widely investigated as alternatives for
humans and the environment. the treatment of secondary wastewater effluent, and in
Some of these adverse environmental effects of phar- the pre-treatment and disinfection step of drinking water
maceuticals are toxicity, development of resistant path- (see Fig. 1 [14]).
ogenic bacteria, genotoxicity and endocrine disruption Also, behavior and elimination of pharmaceuticals
[3,4]. Although many pharmaceuticals do not exhibit have been studied during chlorination, which is a final
acute toxicity, they can have a significant cumulative step of treatment for drinking water and wastewater to
effect on the metabolism of non-target organisms [5] and reduce microbial contamination. Chlorination with
the ecosystem as a whole [6]. Pharmaceutical com- chlorine dioxide and chlorine usually renders both
pounds target specific receptors in humans, whereas chlorination and oxidation products, whereas concerns
many of them are common with other mammals, ver- exist that products with increased toxicity and biological
tebrates and invertebrates. For example, ß-blockers bind activity relative to the parent compound might be gen-
to ß-andregenic receptors in humans and block the ac- erated.
tion of catecholamines (i.e. noradrenalin and adrenalin). For disinfection and oxidation, ozone (O3) is another
However, their effect was seen on fish [7] and inverte- widely used oxidant in drinking water (e.g., for taste and
brate Daphnia magna [8]. More importantly, even when odor control, discoloration, micropollutant elimination).
their toxicity as individual compounds is negligible, they During ozonation, organic compounds either directly
might act in an additive manner in a mixture with other react with O3 in specific reactions, or they decompose
ß-blockers and/or their metabolites exhibiting the same through hydroxyl (OH) radical-mediated reactions.
mode of action, as shown in the test with D. magna [9] Molecular O3 and OH radical have high oxidation
and phototoxicity assays with Vibrio fischeri [10]. For potentials, 2.07 V and 2.80 V, respectively [15]. While
diclofenac, harmful effects were reported on rainbow- O3 is highly selective towards nucleophilic moieties (e.g.,
trout fish [11]. Also, antibiotics present in the environ- carbon-carbon double bonds, aromatic rings, sulfur,
ment could lead to selection of resistant bacterial strains phosphorus, nitrogen and oxygen atoms), OH radical is
[12]. Furthermore, knowledge is greatly lacking on the non-selective and can react with various organic and
possibilities of pharmaceutical residues reaching humans inorganic compounds through hydrogen abstraction,
through biomagnification in the food chain. radical-radical reactions, electrophilic addition, and
Recent studies demonstrated the capability of some electron-transfer reactions [16]. At low or neutral pH
pharmaceuticals (e.g., antidepressants, ß-blockers or li- values, these radical reactions are of minor importance
pid regulators) to bioaccumulate in aquatic organisms regarding degradation reactions of micropollutants. One
[9,13]. Also, all of the abovementioned ecotoxicological possibility to increase the concentration of highly-reac-
effects of pharmaceuticals are of even more concern if tive OH radicals is ozonation at pH > 8, which is con-
their concentration levels in the aquatic environment sidered an AOP [17]. Other AOP examples include O3/
increase (e.g., due to demographic reasons with an H2O2, O3/UV, H2O2/UV, Fenton (Fe2+/H2O2), photo-
increasing percentage of older people, as can be expected Fenton, electro-Fenton, chelating-agent-assisted Fenton/
in the decades ahead). photo-Fenton, heterogeneous photo-oxidation using
WASTE WATER TREATMENT
sewage primary secondary tertiary surface and

system treatment (biological) treatment ground water
treatment
ozonation, AOPs chlorination
DRINKING WATER TREATMENT

coaggulation, distribution
preoxidation filtration,
flocculation, system
disinfection
sedimentation
ozonation,
ozonation, AOPs
AOPs chlorination
Figure 1. Possible applications of ozonation and advanced oxidation processes (AOPs) in treatment of wastewater and drinking water (adapted
from [14]).
http://www.elsevier.com/locate/trac 563
Trends Trends in Analytical Chemistry, Vol. 28, No. 5, 2009
titanium dioxide (TiO2), c radiolysis, and sonolysis [16]. are frequently engaged as complementary analytical
The versatility of AOPs is reflected in different options for tools [e.g., LC-nuclear magnetic resonance (NMR) spec-
producing OH radicals, by combining various chemical troscopy, LC-ultraviolet (UV) detection, LC-infrared (IR)
agents [e.g., O3, hydrogen peroxide (H2O2), transition spectroscopy, GC-MS and LC-MS].
metals, and metal oxides] and auxiliary energy sources Given the importance of accurate identification of
(e.g., UV-vis radiation, electronic current, c-radiation degradation products, we summarize major studies per-
and ultrasound). Hydroxyl radicals are considered the formed with ozonation, chlorination and AOP treatment
primary oxidants in AOPs, while other radical and ac- of pharmaceuticals, including advanced MS methods
tive-oxygen species [e.g., superoxide radical anions and conventional detectors coupled to GC or LC. In
(O2 ), hydroperoxyl radicals (HO2 ), triplet oxygen (3O2), contrast to previously published works on this subject
and organic peroxyl radicals (ROO)] are also involved [22,23], we primarily focus on the combined approach
[16]. of using advanced MS methods for structural identifica-
The main concern about application of these treat- tion of oxidation intermediates of pharmaceuticals, and
ments is formation of various by-products as a conse- bioanalytical assays for evaluation of the toxicity of the
quence of the non-selectivity of OH radicals, that trigger newly identified products.
complex, parallel, consecutive pathways of by-products. For a complete risk-assessment study on the trans-
However, selective reactions of chlorination and formation products (TPs) of pharmaceuticals formed
ozonation have also proved to render persistent and even during treatment of drinking water and wastewater,
toxic intermediates [18,19], so disappearance of the determination of their ecotoxicity is fundamental and a
original drug does not imply that the treatment was prerequisite for comprehensive protection of the envi-
efficient. Some conventional parameters [e.g., total or- ronment. Since only a few published scientific papers
ganic carbon (TOC), chemical oxygen demand (COD), encompass both aspects, we also discuss major studies
dissolved organic carbon (DOC), absorbable organic on the elucidation of unknown products, in order to
halogen (AOX), and aromaticity] can be used to estimate illustrate the advantages and drawbacks of different MS
the process, since, even when possible, complete miner- methodologies.
alization does not seem to be cost effective. However, we do not discuss degradation products of
Considering the potential hazards of by-products of pharmaceuticals formed during biological wastewater
pharmaceutical residues generated during chlorination, treatment (i.e. microbial metabolites), due to a complete
ozonation and AOP, their identification and quantifica- lack of data on their ecotoxicity. Along with biodegra-
tion, as well as elucidation of main reaction mecha- dation processes, pharmaceutical residues can undergo
nisms, are necessary for safe application of such complex enzymatic reactions that can result in microbial
processes for water treatment. These products may have products with toxicity equal to or greater than that of
preserved the mode of action of the parent compound or the original, parent compound [23]. The absence of such
even be biologically more active, hence quantitative studies is of major concern and future investigations
evaluation and determination of their radical rate con- should focus on complete evaluation of the hazards of
stants would afford kinetic and mechanistic data for biotransformation products.
estimating efficiency in removing pharmaceuticals from
real waters that contain high levels of dissolved natural
organic matter (NOM) and other hydroxyl radical scav- 2. GC and LC combined with conventional
engers. Furthermore, in the cases where AOPs (e.g., TiO2 detectors and mass spectrometers
and photo-Fenton photocatalysis) are combined with the
biological treatment, the intermediates formed could be GC-MS has long been a method of choice for identifying
more resistant to degradation by activated sludge and/or volatile compounds in complex mixtures. However, since
inhibit the activity of microbial population [20,21]. products of biological and oxidative transformation are
Novel mass-spectrometry (MS) analyzers [e.g., (quad- often more polar than the original compound, GC-MS
rupole) time-of-flight ((Qq)ToF) enabling accurate-mass techniques frequently cannot cope with identifying them
measurements, linear ion trap (LIT) and hybrid quad- due to their poor volatility. Another major drawback is
rupole LIT (QqLIT) offering high-sensitivity multiple MS that GC-MS is unsuitable for the analysis of thermally-
(MSn) experiments, and LTQ Orbitrap combining both labile compounds. Moreover, the selectiveness of the
features] are powerful tools for identifying so-called derivatization step distinguishes potential new products
‘‘known unknowns’’ (i.e. transformation or degradation with structures different from that of the target com-
products of pharmaceuticals formed during treatment of pound. Nevertheless, Zhang et al. [24] and Andreozzi
drinking water and wastewater). Although these highly- et al. [25] used the full-scan electron impact (EI) mode of
sensitive, selective methods afford reliable results in the GC-MS to identify the reaction intermediates of
process of qualitative analysis, gas chromatography (GC) degradation of acetaminophen by TiO2 photocatalysis,
and liquid chromatography (LC)-hyphenated techniques UV/H2O2 and O3. After derivatization of liquid samples
564 http://www.elsevier.com/locate/trac
(i.e. silylation with BSA or BTSFA+TMCS), hydroxyl and to those obtained for the molecules containing quinaz-
carboxyl groups were converted into their volatile TMS- oline and benzaldehyde moieties. These chemical shifts
ether and TMS-ester derivatives. exhibited great similarities, so the hypothesized structure
GC-MS is frequently employed for detection of apolar of BQD was finally confirmed. Nevertheless, ozonation
degradation products, and/or as an identification tool products BQM and BaQD could not be isolated for NMR
complementary to LC-MS [26,27]. For example, measurements. The need to separate individual com-
Lambropoulou et al. [26] detected products of TiO2 pounds without impurities represents a major drawback
photocatalytic degradation of bezafibrate 4-isopropoxy- of NMR analysis, since in processes, such as chlorina-
benzaldehyde, 4-chloro-benzaldehyde and 4-chloro- tion, ozonation and AOPs, multiple degradation products
benzamide by fitting the recorded EI mass spectra using will be formed.
the NIST library. Of these three intermediates, only 4- Buth et al. [29] successfully combined 1H-NMR and
chloro-benzamide was detectable by LC-ToF-MS, proba- 2D-NMR spectroscopy with high-performance LC
bly due to the low ionization efficiency, highly polar (HPLC)-ToF MS operated in electrospray ionization (ESI)
character and/or thermal instability and low volatility of mode and IR spectroscopy to characterize products iso-
the other two compounds. Pérez-Estrada et al. [27] lated after chlorination of cimetidine. Mechanisms of
managed to identify apolar degradates of diclofenac chlorination were determined to proceed via cimetidine
formed in the photo-Fenton treatment by employing GC- sulfoxide as the primary intermediate, rendering the
(EI)-MS. Furthermore, analysis in positive chemical same terminal products [i.e. 4-hydroxymethyl-5-methyl-
ionization (PCI) mode with methane as the reagent gas 1H-imidazole, 4-chloro-5-methyl-1H-imidazole, and
enabled accurate confirmations of the molecular weights either a ß-sultam (N-cyano-NÕ-methyl-NÕÕ-ß-sultamyl-
and detection of co-eluting compounds. The intermedi- guanidine) or a d-sultam (N-(2-methyl-1,1-dioxide-
ate products identified did not pose the quinone imine 1,2,4-thiadizinan-3-ylidene) cyanamide)], independently
group as determined for the primary degradation route of pH. The first two products (i.e. cimetidine sulfoxide
of diclofenac, yet they mainly resulted from direct reac- and 4-hydroxymethyl-5-methyl-1H-imidazole) were
tion of the aliphatic chain in the structure of the com- identified by matching their HPLC retention times with
pound. the authentic standard and 1H-NMR spectra. Further, 4-
In ozonation experiments with carbamazepine, chloro-5-methyl-1H-imidazole revealed a monochlorine
McDowell et al. [28] identified three TPs of carbamaze- isotopic pattern and 1H-13C correlations in positions very
pine [i.e. 1-(2-benzaldehyde)-4-hydro-(1H,3H)-quinazo- similar to those for the methyl and 2-position protons
line-2-one (BQM), 1-(2-benzaldehyde)-(1H,3H)-quinaz- seen in the spectrum of 4-hydroxymethyl-5-methyl-1H-
oline-2,4-dione (BQD) and 1-(2-benzoic acid)-(1H,3H)- imidazole, obtained in 2-D heteronuclear multiple-bond
quinazoline-2,4-dione (BaQD)]. Oxidation-product BQD correlation (HBMC) experiments using NMR spectros-
was sufficiently apolar and volatile to yield its GC-(EI)- copy. However, the methylene signal missing in the
MS spectrum, whereas additional fragmentation experi- spectrum of chlorinated derivative indicated replacement
ments were performed on GC-ion trap (IT)-MS. The of the hydroxylmethyl group with chlorine. Finally, the
fragmentation pattern of BQD was compared to frag- fourth product could not be unequivocally identified as
mentations of comparable quinazoline compounds [i.e. ß-sultam or d-sultam by the available spectrometric and
3-methyl-(1H,3H)-quinazoline-2,4-dione, a sex phero- spectroscopic techniques. The peak recorded in ESI(+)
mone in the pale-brown chafer, and caffeine]. The major mode corresponded to the sodium adduct, [M+Na]+
fragmentations of these compounds via a two-bond 211.0244, which, after H/D-exchange (use of D2O as
cleavage on their heterocyclic rings suggested nitrogen- HPLC solvent instead of water) appeared at m/z 214.0.
driven and carbonyl-oxygen-driven a-cleavages at the The three acid protons exchanged were assigned to the
hetero-ring during ozonation of carbamazepine. How- proton on the secondary amine and two protons at the
ever, the structure of BQD could not be unambiguously a-C-atom to the sulfonamide functionality. When
elucidated from the MS data. In order to gain more employing chemical ionization (CI) instead of ESI, a
information, other techniques were used (e.g., 1H NMR, fragment ion at m/z 64 was detected, probably resulting
COSY, 13C NMR and spin-echo experiments). from the loss of SO2. Further evidence for the sultam
NMR spectroscopy (in the first place 1H-NMR, 13C- structure of this intermediate was provided by its 1H
NMR and 15N-NMR) is often combined with other GC- NMR and IR spectra, although without the possibility of
MS and LC-MS methods in qualitative analysis of phar- distinguishing between the ß isomer and the d isomer.
maceuticals. Frequently, the 1H-NMR spectrum cannot Nevertheless, this product was somewhat unexpected as
provide unequivocal evidence of the structure, as was the outcome of chlorination, since it resulted from C-S
the case for the BQD-ozonation product of carbamaze- cleavage, intramolecular nucleophilic substitution and
pine. The 13C-NMR spectrum enabled assignments of oxidation.
chemical shifts produced from 13C signals of the products Another study from Dodd and Huang [30] illustrated
dissolved in methylene chloride-D2 and compared them the importance of combining supplementary analytical
techniques. In the reaction of sulfamethoxazole with free products obtained by chlorination of the aromatic ring,
chlorine, degradation product N-chloro-p-benzoquin- probably in the ortho position to the phenol group, thus
oneimine (NCBQ) with potentially higher acute toxicity affording chloro-4-acetamidophenol. Dichlorinated
than the parent compound was detected by 1H-NMR and acetaminophen was also recognized from its character-
GC-MS, yet it yielded no signal in positive ESI or atmo- istic isotopic pattern with [M+H]+ 220.0 and additional
spheric pressure chemical ionization (APCI) modes of LC- ions at m/z 222.0 and 224.1, whereas ammonium and
MS analysis. sodium adducts were detected at m/z 237.0 and 239.0.
Secondary oxidation products of pharmaceuticals
(e.g., phenols and acids) can usually be identified by
simple detectors [e.g., UV, fluorescence (FLD) and diode 3. LC-MS methods based on MSn experiments
array (DAD)]. Using HPLC-DAD and HPLC-FLD, Doll
et al. [31] identified 4-chlorophenol, hydroquinone and Having a structure similar to that of the parent com-
isobutyric acid as degradation products of clofibric acid pound, TPs are not complete unknowns. Their structures
in TiO2 suspensions. However, primary degradation are elucidated by comparing mass spectra of analyte and
products 2-(4-hydroxyphenoxy)-isobutyric acid (m/z its TPs, considering possible molecular changes relative
195), hydroxyisobutyric acid (m/z 103) and 4-chloro- to the parent compound and using putative transfor-
hydroxyphenol (m/z 143, 145, ratio 3:1) were tenta- mation pathways. To facilitate the identification process,
tively identified by HPLC-triple quadrupole (QqQ)-MS. derivatization of functional groups of the products
In another study from the same authors [32], the TiO2 formed and/or H/D-exchange experiments can also be
photocatalytic degradation product of carbamazepine – employed [34,35]. Nevertheless, lack of analytical ref-
10,11-dihydro-carbamazepine-10,11-epoxide – was ide- erence standards means that accurate quantification of
ntified by its LC retention time and UV spectrum, based degradation products of pharmaceuticals is rarely pos-
on comparison with authentic standards. Other degra- sible.
dates (i.e. hydroxycarbamazepine and dihydroxycarb- IT-MS has proved to be a very useful analytical tool,
amazepine, acridine derivatives) for which no authentic due to its ability to determine multi-stage fragmentation
standard was available were characterized by their pathways of molecules through MSn experiments. In the
tandem MS (MS2) spectra obtained using LC-QqQ-MS. trap comprising two hemispherical electrodes and a ring
For determination of chlorination products of acet- electrode, IT-MS can trap and accumulate ions that are
aminophen, Bedner et al. [18] engaged two LC-UV sys- further focused toward the centre of the trap by using an
tems, one with electrochemical (EC) and other with MS oscillating voltage. This increases the signal-to-noise
detection. In the full-scan performed in ESI(+) mode at (S/N) ratio, although it entails some drawbacks (e.g.,
LC-UV-MS, product N-acetyl-benzoquinone-imine small trapping volume, loss of the dynamic range re-
(NAPQI) emerged as a methanol adduct [M+H+MeOH]+ sponse due to space charging, and overfilling of the IT),
182.1, while the molecular ion [M+H]+ 150.1 was less which result in deterioration in the mass spectrum [36].
intense. This spectrum matched that obtained for a The major advantages of LIT over IT are larger ion-
standard of NAPQI, which, together with the equivalent storage capacity and higher trapping efficiency. The
retention times obtained at LC-UV-EC for the authentic QqLIT (i.e. QTRAP-MS) offers an additional possibility to
standard and the sample, confirmed the identity of this operate the third quadrupole as a normal quadrupole or
product. The mass spectrum of the second degradation in the LIT mode.
product of acetaminophen – 1,4-benzoquinone – could The advantage of the MSn capability of LIT-MS was
not be obtained in the ESI mode, whereas, in the APCI nicely illustrated in studies on the identification of
mode, one ion was seen in the spectra of both sample products of ozonation, TiO2 and UV/TiO2 photocatalysis
and the standard of 1,4-benzoquinone, at m/z 108.1. of ciprofloxacin [37,38], and TiO2 photocatalysis of
Since the nominal mass of 1,4-benzoquione is 108.0, the analgesic anxiolytic drug buspirone [39] and anti-
ion was explained by either electron capture to form a inflammatory drug diclofenac [40]. Paul et al. [37] used
radical anion or by a mass gain that was exactly bal- LC-(ESI+)-MS2 analysis with an IT instrument to identify
anced by a mass loss to make a negative ion. Also, the six degradation products of UV and visible-light TiO2
retention times recorded by LC-UV-EC for the sample and photocatalytic degradation of ciprofloxacin; all six deg-
the standard were identical. Furthermore, the most radation products retained the oxoquinoline carboxylic
prominent product in the spectrum was detected with acid group. Also, De Witte et al. [38] determined the
the molecular ion [M+H]+ 186.1, which exhibited a ozonation-degradation pathway of ciprofloxacin by LC-
monochlorinated pattern with the second ion appearing (UV)-IT-MS operating in low-resolution MS (LRMS) and
at m/z 188.1. The combination of EC detection with io- high-resolution MS (HRMS) modes. Collision-induced
dide-post-column reaction chemistry [33] allowed dissociation (CID)-MS fragmentation revealed charac-
exclusion of any possible chloramides, and that sug- teristic losses of isatin analogues (i.e. CO and C2O2),
gested that the [M+H]+ 186.1 corresponded to the indicating the presence of a diketone. However, in the
MS2 spectra obtained for anthranilic-acid analogues, imine transformation. Two dihydroxylated derivatives
losses of H2O, H2O and CO, and H2O and C2O2 were with [M+H]+ 328 and [M+H]+ 284 were also identified,
recorded, whereas the hypothesized structures were generated by OH radical attack on hydroxyl-diclofenac
tentatively identified by comparison with the spectrum of and consequent decarboxylation, respectively.
analogous N-formylanthranilic acid. Whereas isatin and Although the usual operating mode of MS analyzers is
anthranilic-acid analogues were derived from degrada- ESI due to enhanced detection of polar compounds,
tion of the quinolone moiety of ciprofloxacin, oxidation several studies [45,46] used the APCI interface, which is
reactions also occurred at the piperazinyl fraction, with suitable for less polar, thermally inert compounds. Calza
the main degradation product formed through a net loss et al. [45] identified hydroxylated and R-NH2 derivatives
of C2H2, desethylene ciprofloxacin. Desethylene cipro- formed in TiO2-photocatalytic degradation of sulfon-
floxacin was univocally identified based on nominal amide drugs (i.e. sulfadiazine, sulfamerazine, sulfadime-
mass and HPLC retention time. All products were com- thoxine and sulfathiazole) by MS2 experiments
pletely mineralized during UV photocatalysis, whereas, performed on an LC-(APCI)-IT-MS instrument. More-
in the visible-light assisted process, the fluoroquinolone- over, one of the intermediates identified for sulfamera-
core structure was possibly preserved. However, whether zine – 4-methyl-2-aminopyrimidine – was stable in the
the degradation of ciprofloxacin will occur primarily at degradation time investigated (i.e. 30 min). Further-
the carboxylic group (quinolone moiety) with formation more, APCI was used to identify ozonation intermediates
of isatin and anthranilic acid analogues, or at the pip- of tetracycline, because typical conditions of this ioni-
erazinyl substituent will depend on the pH of the solu- zation can prevent interface obstruction usually caused
tion. In the first case, at neutral pH, carboxylic and keto by the required use of oxalic acid (a non-volatile com-
groups, which are considered necessary for binding pound) [46].
quinolones to the DNA gyrase target, are degraded [41].
However, if the ozonation is conducted at pH 3 or pH 10,
intermediates formed preserve the essential structure of 4. LC-MS methods based on accurate-mass
antibacterial quinolone, whereas the data on the activity measurements
of desethylene ciprofloxacin reported in literature are still
contradictory [42–44]. This stresses the importance of ToF-MS represents an indispensable analytical tool for
optimizing ozonation in treating water loaded with the non-target screening and structural elucidation of
quinolones. metabolites and TPs, due to its full-scan sensitivity, high
Calza et al. [39] identified several intermediates of selectivity (lack of interferences) and specificity (correct
TiO2-photocatalytic degradation of buspirone, which empirical formula assigned), given by the high mass
were also found in the in vivo experiments on rats and resolution and mass accuracy of ToF analyzers. It is
horses. Buspirone and its degradation products were based on the acceleration of ions by an electric field,
characterized by MSn experiments, conducted on a LC- whereas the time that ions need to reach the detector
(ESI)-IT-MS instrument, by linking the main fragments will depend on their kinetic energy, and thus their m/z
in the fragmentation pathway for buspirone shown in ratios. When a ToF-MS is interfaced with commonly
Fig. 2. In particular, the product ion at m/z 122 will be used continuous ion sources (e.g., ESI), the ions are
indicative of the unchanged pyrimidine substructure, introduced into the mass analyzer in a direction per-
while the m/z 222 and m/z 180 ions are diagnostic of the pendicular to the direction of flight and accelerated down
unchanged azaspirone-decane-dione substructure. the ToF flight tube in a pulsed fashion. This orthogonal
In another study published by the same authors [40], accelerating ToF (oaToF)-MS has high mass accuracy
the same methodology was used to identify several hy- (generally within a few parts per million (ppm) of the
droxyl and dihydroxy derivatives formed during TiO2 exact m/z values calculated from the nuclide masses and
photocatalytic treatment of diclofenac, which further the ionic charge z), resolution of 10 000–20 000
transformed into chlorophenol or hydroxylphenol prod- (FWHM) and allows rapid mass scanning in a theoreti-
ucts. Four products holding molecular ion [M+H]+ 312 cally unlimited mass range [47]. In a QqToF instrument,
were detected, exhibiting the same characteristic losses the precursor ion is selected in a quadrupole mass fitter
of water, formic acid and joint losses of formic acid and and dissociated in a radiofrequency (RF)-only multipole
chlorine radical as those observed in the fragmentation collision cell, and the resultant fragments are analyzed in
pathway of diclofenac. It was therefore deduced that the a ToF-MS instrument. A CID spectrum with accurate
hydroxylation took place at the aromatic rings. The masses of both precursor and product ions is therefore
discrepancies in the MS3 spectrum of the fourth product impossible with LRMS analyzers. However, one of the
with [M+H]+ 312 relative to the above-mentioned greatest drawbacks of this instrument is low ion trans-
compounds situated the hydroxyl group at the lateral mission (resulting in poor MS2 sensitivity and detection
chain of diclofenac. A product having 2 Da less than the limits) and a limited intensity range over which accu-
monohydroxylated derivatives pointed to a quinine- rate-mass data can be acquired.
Figure 2. Fragmentation pathway followed by buspirone (reproduced from [39] with permission of Elsevier).
In our own works, we identified intermediate products paring the fragments of P285 with that of ranitidine,
of TiO2 and photo-Fenton treatment of ß-blocker aten- mutual loss of 46 Da was detected, corresponding to the
olol and anti-histaminic drug ranitidine solely by ultra- loss of the NO2 radical. The rupture of the C-S bond led to
performance LC (UPLC)-(ESI+)-QqToF-MS [48,49]. We the formation of a methylated furan-aldehyde fragment
elucidated the structures and interpreted the fragmen- ion at m/z 109, similar to the fragment ion m/z 138
tation pathways of atenolol, ranitidine and the detected formed in the case of ranitidine. A degradation product
TPs acquired in CID MS2 experiments under optimized carrying a molecular ion [M+H]+ 270.0900 was then
conditions of collision energy (CE) and cone voltage (CV). assumed to be formed by photocatalytic cleavage of
The unequivocal identification of the degradation prod- oxygen from the nitrite group (-NO2) in product P285,
ucts was facilitated by the MassLynx V4.1 software and that was confirmed by the absence of the charac-
identification program of the QqToF instrument with a teristic loss of 46 Da. The absence of typical loss of N,N-
mass-measurement-accuracy threshold of 5 ppm. Inter- dimethyl amino moiety in product P300 ([M+H]+
pretation of the fragmentation pathways of the com- 301.1362) suggested cleavage of the terminal methyl
pounds detected was mainly supported by the analogies group in the tertiary amine function. Next, three more
and the differences with the MS2 spectra of original peaks detected in the full-scan mode of QqToF instru-
compounds and the oxidation reactions expected in the ment were assigned to hydroxylated derivatives of rani-
processes investigated. tidine and their quinone-imine analogues. These
In photocatalytic experiments with ranitidine [48], a products had molecular ions [M+H]+ 331.1423,
product having a molecular ion [M+H]+ 286, denomi- 345.1266 and 361.1145, which corresponded to one
nated as P285, was hypothesized to be formed by con- more oxygen atom, two more oxygen and two less
secutive cleavages of the N,N-dimethyl amino moiety hydrogen atoms, and three more oxygen and two less
and OH radical attack on the terminal methyl group hydrogen atoms relative to ranitidine, respectively. Their
substituent, followed by addition of oxygen. By com- peculiar MS2 fragments were linked together in a frag-
mentation pathway, whereas both fragment ions and bond was observed for all intermediates encountered,
radical fragment ions were confirmed by rational and indicating an intact isopropyl moiety (see Fig. 3).
integer double-bond equivalent (DBE) values, respec- In several studies, exact-mass assignment to TPs
tively. available with ToF-MS was combined with GC-(EI)-MS in
The same approach was applied when elucidating the order to determine the entire degradation pathways of
structures of degradation products of atenolol [49] (see pharmaceuticals, including polar, less polar and/or
Fig. 3). The difference in mass of 29 Da between product apolar products [26,27,50]. Pérez-Estrada et al. [50]
P237 ([M+H]+ 238.1444) compared with the molecular elucidated structures of several degradation products of
ion of atenolol ([M+H]+ 267.1702) was assumed to be analgesic drug dipyrone, formed during TiO2-assisted
due to the loss of an amide group from the molecule and and photo-Fenton photocatalysis. GC-MS analysis en-
addition of oxygen to the alkyl group attached to the abled identification of five intermediates, which were
ring after hydrogen abstraction by OH radical attack, tentatively identified by their full-scan mass spectra.
resulting in a keto derivative. In the MS2 spectrum of Four products were formed by an OH radical attack on
both compounds, characteristic losses of isopropyl and the double bond in the heterocyclic ring and had some
amino isopropyl groups and water were observed. Two common fragments (m/z 121, 107, 92 and 77). The fifth
intermediates with molecular ions [M+H]+ 283.1656 product detected by GC-MS preserved the heterocyclic
(P282) and 281. 1505 (P280) corresponded to a ring, while the other five degradation products having
monohydroxylated atenolol and its quinone-imine the heterocyclic ring were identified by LC-(ESI)-ToF-MS.
derivative, respectively. Again, a characteristic loss of Since there was no possibility of performing CID experi-
the isopropyl group (42 Da) in the spectra of P280 and ments, the proposals for chemical structures of these
282 afforded fragment ions m/z 239 and 241, respec- compounds were based on their accurate masses mea-
tively. These fragments further exhibited consequent sured for the molecular ions (error <5 ppm), oxidative
losses of acetamide moiety, water and ammonia. The process chemistry and DBE data given by the software.
fragment ion at m/z 133.028 in the fragmentation pat- Moreover, the amounts of nitrate and ammonia deter-
tern of P280 was possibly formed by the loss of the mined by ion chromatography (IC) provided additional
methyl-isopropylamine group, water and the acetamide information for hypothesizing a possible oxidation
moiety from the molecular ion, and intramolecular pathway of dipyrone.
cyclization afterwards. Besides these products, at reten- In another study from the same authors [27], the
tion times 2.5 min, 4.3 min and 5.15 min, three peaks combination of GC-MS and LC-ToF-MS was engaged in
with the same molecular ion [M+H]+ 254 were detected. identifying the photo-Fenton intermediates of diclofenac.
Since they presented the same losses in the mass spectra Again, apolar degradation products were determined
with the same relative ion-signal intensities, they were based on their full-scan EI spectra matching with spec-
all labeled as P253, whereas the non-specific OH radical tral libraries or by their fragmentation patterns, whereas
attack on the P237 intermediate was assumed to be semi-polar and polar products were analyzed by HPLC-
responsible for their formation. Since there were only (ESI)-ToF-MS. The two early appearing peaks with
two available positions for the addition of the OH radi- molecular ions [M+H]+ 312.0193 and 310.0031 cor-
cal, ortho- and meta-, to the aldehyde group, three dif- responded to one more oxygen, and one more oxygen
ferent peaks could be explained by the intramolecular and two hydrogen less compared to the diclofenac mol-
hydrogen bond between the hydrogen of the –OH group ecule, respectively. This suggested that a hydroxyl group
attached to the ring and oxygen of the aldehyde group was attached to the aromatic ring, whereas the second
(ortho isomer) or the absence of one (meta isomer). It product corresponded to its quinone-imine derivative.
could be assumed that, in the case of the meta isomer, Based on the measured masses of protonated molecules,
the hydrogen bond with the ether oxygen from the alkyl as well as frontier electron density (FED) theory [which
backbone would be less probable since the electron predicts sites of OH radical attack by calculating the
density at ether oxygen is lowered by the alkyl sub- higher molecular orbital (HOMO)], formulae for other
stituent. The presence of intramolecular hydrogen bond intermediate products were proposed, while the resulting
has been previously reported for ortho isomers of accurate-mass errors were always <2.5 ppm. The
hydroxylated photocatalytic intermediates of bezafibrate accuracy of the ToF instrument was enhanced by
[26], whereas an increase in retention time was ob- internal calibration, whereby internal references were
served due to the decrease in polarity of the molecule. introduced with a dual-nebulizer ion source at a very
Thus, the most polar compound eluting at 2.5 min was low flow rate, with constant autocalibration and
assigned to meta-hydroxylated intermediate P253, while recording of reference mass along with the raw data.
the compounds appearing at 4.3 min and 5.15 min were Lambropoulou et al. [26] used GC-MS, HPLC-ToF-MS
possibly the two ortho derivatives, without and with the and HPLC-DAD to identify up to 17 degradation prod-
intramolecular hydrogen bond, respectively. Finally, the ucts of TiO2 photocatalysis of bezafibrate under simu-
m/z 116 fragment ion, obtained by breaking the C-O lated solar light. LC-ToF analysis was able to detect
145.0653
100
a) NH2 H
O
N O
H
[M+H-C(CH3)2-NH3-H2O]+
OH
190.0869
NH2
%
[M+H]+
N O
H
O O 267.1702
OH
[M+H-C(CH3)2-NH3]+
OH
208.0974
133.0660
[M+H-C(CH3)2]+
116.1065 225.1240
m/z
0
100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290
238.1444
100
b) O
O O
[M+H]+
O
[M+H-C(CH3)2]+
%
161.0608 196.0965
N
H
133.0654 149.0607 [M+H-C(CH3)2-H2O]+ [M+H-H2O]+

116.1065 178.0861 220.1336
m/z
0
100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250
281.1505
100
c) O
O [M+H]+
N O
O H
OH
[M+H-C(CH3)2-NH3-CO]+
O
194.0822 O
%
O
O O
O
O
OH
116.1066 121.0257 [M+H-C(CH3)2]+

133.0282 176.0713 239.1049
159.0440
m/z
0
100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290
178.0861
100 NH2
d) OH
O
[M+H-C(CH3)2-NH3-H2O-CO]+ N
H
O
[M+H]+
OH 283.1656
[M+H-C(CH3)2-NH3-H2O]+
OH
O
O
161.0595 O
%
116.1069 OH 206.0820 OH
O 224.0935
[M+H-C(CH3)2]+
149.0594 241.1163
196.0972 [M+H-NH3]+
266.1389
m/z
0
100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290
Figure 3. Spectra obtained in ESI(+)-MS2 experiments on a QqToF instrument (cone voltages 25 V, collision energies 15–25 eV) for atenolol and
the intermediate products of its photocatalytic degradation: a) atenolol; b) P237; c) P281; and, d) P283 (reproduced from [48] with permission of
Elsevier).
various hydroxylated products in the first stage of the R-oxy group of bezafibrate on the aromatic ring, these
treatment, based on accurate-mass measurement of two products were identified as monohydroxylated
protonated molecules, abundances of the chlorine iso- derivatives (meta- and ortho- at 19.7 min and 20.9 min,
tope in the molecular ions, characteristic fragment ions respectively). In this case, consideration of their polarity,
used as diagnostic ions, and, when possible, accurate- which determined the order of elution, helped assign the
mass measurements of sodium adducts of molecular peaks to their corresponding isomers. Namely, the ortho
ions. A fragment ion from the spectrum of bezafibrate isomer assumed an intramolecular hydrogen bond be-
(Cl35 [M+H]+ 362.1154, Cl37 [M+H]+ 364.1128) tween the hydrogen of the hydroxyl group borne by the
exhibiting a chlorine-isotope pattern at m/z 138.9965 ring and the oxygen of the ether group.
and 140.99 was used for diagnosis. The software tool The latest advances in MS are LTQ Orbitrap and
calculated a unique elemental composition correspond- Fourier transform (FT)-MS. Next-generation FT ion
ing to this diagnostic ion, C7H3ClO, whose structure is cyclotron resonance (FT-ICR)-MS accumulates ions
shown in the inset in Fig. 4. Targeting the m/z of the externally to a superconducting magnet, and, unlike
diagnostic ion, numerous chlorine-containing products other MS techniques where ions are detected by hitting a
were found in the total ion chromatogram (TIC). Several detector (e.g., electron multiplier), in FT-ICR, ions only
monohydroxylated products were detected, as a conse- pass near detection plates, and the m/z ratio of ions is
quence of OH radical attack on the 4-chlorobenzoyl ring determined based on the cyclotron frequency of the ions
or the phenoxy moiety. Interestingly, two peaks that in a fixed magnetic field [51]. Besides the high trapping
eluted prior to bezafibrate at 19.7 min and 20.9 min had capacity and MSn capabilities, FT-ICR has the automatic
identical mass spectra and the same molecular ion gain control (AGC) of LIT-MS with unsurpassed mass
[M+H]+ 378.1100. From the best-fit formula given by accuracy, fast data collection, good sensitivity, good
the software and considering the activating effect of the dynamic range and high resolution, although the high
Figure 4. Total ion chromatogram (TIC) by LC-ToF-MS obtained from an SPE extract of bezafibrate solution after 60 min of irradiation with sim-
ulated solar light in the presence of TiO2 suspension (100 mg/L). The TIC shows the protonated molecular ion of bezafibrate at m/z 362 and
chlorinated by-products. Identification criteria for chlorine by-products: (i) searching of diagnostic ion-accurate mass 138.9965 (remaining frag-
ment from bezafibrate); and, (ii) chlorine isotopic profile (reproduced from [26] with permission of Elsevier).
cost of these instruments limits wider application in hydroxyphenylacetamide mutual with the parent com-
environmental and pharmaceutical laboratories [52]. pound, unusual loss of methanol was seen, suggesting
Due to the high cost of FT-ICR spectrometers, LTQ that one of the methyl groups was hydroxylated. Also,
Orbitrap is considered a good alternative because of its two other compounds, having molecular ions [M+H]+
high resolution and good mass accuracy [52]. The LTQ 283.1658, exhibited a fragmentation pattern similar to
Orbitrap is a hybrid MS combining LIT and Orbitrap, that recorded for atenolol, but with m/z values of frag-
enabling excellent mass resolution and accuracy, as well ments 15.9949 units higher. This implied insertion of a
as high sensitivity. The ions are introduced tangentially hydroxyl group into the aromatic ring at two positions,
into the electric field between the inner spindle-like which afforded two isomers of monohydroxylated
electrode (central electrode) and an outer barrel-like, derivatives. Another pair of isomers with molecular ions
coaxial electrode, and trapped because their electrostatic [M+H]+ 283.1658 was formed by attaching the hydro-
attraction to the inner electrode is balanced by centrif- xyl group on the side-chain of atenolol, which exhibited
ugal forces. Ions of a specific m/z ratio therefore move in the same characteristic losses of propene, isopropylamine
rings that oscillate along the central spindle, while the and N-(1-methyl-2-hydroxyethyl)-N-(2-hydroxyethyl)-
frequency of these harmonic oscillations is independent amine, whereas the latter loss excluded the OH attack
of the ion velocity and inversely proportional to the on the aromatic ring. Since dihydroxylated oxidation
square root of the m/z ratio. products of atenolol were formed in small quantities,
As demonstrated by recent publications [53,54], the only MS2 fragmentation of this compound could be
LTQ Orbitrap mass analyzer supports qualitative analysis performed on the LTQ Orbitrap, which impeded accurate
of degradation products of pharmaceuticals present at attribution of OH radical-attack sites. Finally, six tri-
trace-level concentrations. Seven intermediate products hydroxylated intermediate products with molecular ions
of salbutamol in TiO2-assisted photocatalysis were [M+H]+ 315.1551 were detected, while their identifica-
structurally elucidated by LC-LTQ Orbitrap-MS, by tion was facilitated by analogies with the MSn spectra of
taking into account their retention times, kinetics of the monohydroxylated and dihydroxylated derivatives
evolution and exact-mass measurements of the MS and and atenolol, and some characteristic losses (e.g., one
MSn spectra [54]. Characteristic fragment ions in the loss of water indicated attachment of an –OH group into
spectrum of salbutamol ([M+H]+ 240.1594) were ob- the aromatic ring and loss of allyl alcohol insertion of an
tained by the loss of water (m/z 222.1489), tert-butene –OH group into the propyl chain).
(m/z 166.0863) and again water (m/z 148.0757). The
species having molecular ions [M+H]+ 210.1489 and
226.1438 exhibited the same tert-butene and water 5. Toxicity in degradation/transformation studies
losses, but the characteristic loss of hydroxylmethyl
group was absent. These products were identified as 2- In published studies about TPs of pharmaceuticals, tox-
(tert-butylamino)-1-(3,4-dihydroxyphenyl)ethanol and icity was generally taken as the overall toxicity of the
2-(tert-butylamino)-1-(4-hydroxyohenyl)ethanol. Two solution (i.e. mixture of intermediates and possibly
species with the same nominal mass were identified as parent compound) by the Microtox test (see Table 1)
molecular ions [M+H]+ 224.1654 and 224.1281, [40,50,53–57]. The Microtox test measures the decrease
products of the cleavage of one hydroxyl group and in bioluminescence of the marine bacterium V. fischeri,
methylic group, respectively. The first, [M+H]+ as a result of inhibiting bacterial luciferase upon contact
224.1654, presented the same peculiar losses of tert- with toxic substances. This test is frequently used in the
butene and water, whereas the second, [M+H]+ environmental studies, since it is simple, fast, sensitive
224.1281, followed a different fragmentation pathway and reproducible.
and had a shorter retention time than the parent com- Calza et al. [40,57] followed the evolution of the tox-
pound, in accordance with the loss of a methylic group. icity of solution of diclofenac and imipramine during
Three more products were observed, having molecular TiO2-photocatalytic treatment by using the Microtox
ions [M+H]+ 182.1176, 118.1226 and 132.1019. bioassay. In both studies, there was a maximum inhi-
Medana et al. [53] used an LTQ Orbitrap for struc- bition (expressed as % of inhibition of bioluminescence of
turally elucidating products of the photocatalytic trans- bacteria) of 72% (after 20 min) and 83% (after 15 min)
formation of atenolol. Six species with molecular ions for diclofenac and imipramine solution, respectively.
[M+H]+ 283.1658 were detected, corresponding to These maxima corresponded to the maximum concen-
monohydroxylated derivatives of atenolol as a conse- trations of identified TPs, thus indicating that interme-
quence of non-selective OH radical attack. The exact diates more toxic than imipramine were formed,
positions of the attacks were determined by comparing although it was not possible to identify their individual
their fragmentation pathways to that elucidated for toxicities. In the following period, the observed toxicities
atenolol. For example, for one of the monohydroxylated of the solutions decreased, yet remained at values higher
derivatives, besides neutral losses of water and para- than the initial toxicities of diclofenac and imipramine.
Table 1. Summary of studies that combine qualitative analysis of degradation products of pharmaceuticals in chlorination, ozonation and advanced oxidation
processes (AOPs) with evaluation of the toxicity
Pharmaceutical, Transformation products Analytical method Toxicity Toxicity test Ref.

treatment
Acetaminophen1, NAPQI2, 1,4-benzoquinone3 HPLC/UV/EC, 500 mg/kg1 Intraperitoneal [18]

chlorination mono- and dichlorinated derivatives HPLC/UV/MS 20 mg/kg2 injection in mouse
8.5 mg/kg3
Cimetidine1, Cimetidine sulfoxide2 HPLC-(ESI)-ToF 35 lg/L1,a ECOSAR (daphnid [29]
Chlorination 4-hydroxymethyl-5-methyl-1H-imidazole3 MS, 1H, 2D-NMR, IR 370 lg/L2,b or green algae)
4-chloro-5-methyl-1H-imidazole4 20 lg/L3,a
b-sultam5 6.5 lg/L4,a
d-sultam5 630 lg/L5,b
Diclofenac, mono- and dihydroxylated derivatives, 5- HPLC-IT-MS 72%c max. Microtox [40]
TiO2 photocatalysis hydroxy diclofenac and its keto derivative, (after 20 min) (Vibrio fischeri)
chloro- and hydroxyphenol derivatives 2,6-
dichlorophenol, 4-chlorocathecol
Dipyrone (MAA), TiO2 4-aminoantipyrin, aniline, derivatives of GC-MS, <50% c (TiO2) Microtox [50]
photocatalysis pyrazole ring opening HPLC-(ESI)-ToF-MS 55-60% c (Vibrio fischeri)
Photo-Fenton photocat. (photo-Fenton)
Atenolol, mono- and dihydroxylated derviatives HPLC-LTQ Orbitrap -15% c Microtox [53]
TiO2 photocatalysis (hydroxylation of benzene ring and alkyl (15-60 min) (Vibrio fischeri)
side-chain)
Salbutamol, 2-(t-butylamino)-1-(4-hydroxylphenyl) HPLC-LTQ Orbitrap 54% c Microtox [54]
TiO2 photocatalysis ethanol, (after 15 min) (Vibrio fischeri)
2-(t-butylamino)-1-(3,4-dihydroxyphenyl)
ethanol,
2-(t-butylamino)-1-(4-hydroxyl-3-
methylphenyl) ethanol,
2-(t-butylamino)-1-(3,4-dihydroxyphenyl)
ethanone,
2-(methylamino)-1-(3,4-dihydroxyphenyl)
ethanone
2-(t-butylamino)-ethanol,
2-(t-butylamino)-acetic acid
Ibuprofen, monohydroxylated derivatives (hydroxylation HPLC-(ESI)-ToF-MS 92%c (after 120 min Microtox [56]
TiO2 photocatalysis of methylpropyl phenyl and arylcarboxylic for 2-8 mg/L O2; and (Vibrio fischeri)
moiety) 240 min for 40 mg/L O2)
Imipramine, mono-and polyhydroxylated derivatives, GC-MS, HPLC-(ESI) 83%c (after 15 min) Microtox [57]
TiO2 photocatalysis derviatives with cleaved aminoalkylic side- QqQ-MS (Vibrio fischeri)
chain
Bezafibrate, derivatives of benzene ring opening, HPLC-MS 0.09d max. Microtox [55]
Ozonation trihydroxylated products, derivatives with (after 5-7 min) (Vibrio fischeri)
cleaved non-chlorinated benzene ring
Nitrofurazone1, NFA2, NFOA3, HMF, HFA, chlorinated HPLC-UV 1921,e, 2.642,e, 0.143,e Ames test [70]
Chlorination products (Salmonella
Typhimurium TA100,
without S9 mix)
Sulfamethoxazole, derivative of C-N bond cleavage, methanol, GC-MS, FTIR changes in cellÕs mammalian cell [77]
Ozonation ethanol, phenol morphology line (HepG2 cells)
Clarithoromycin, clarithromycin-N-oxide HPLC-(ESI)-MS2, 25 % of EC50 of P.putida growth [83]
1
ozonation demethylated clarithromycin H-NMR clarithromycin inhibition test
deaminated clarithromycin
acetalized clarithromycin
a
PNEC derived from LC50.
b
PNEC derived from EC50.
c
% of inhibition Vibr. Fischeri.
d
1/EC50,15min.
e
net reverants/nmol. NAPQI- N-acetyl-benzoquinone-imine, MAA- 4-methylaminoantipiryne (hydrolytic product of dipyrone), NFA- 5-nitro-2-
furaldehyde, NFOA- 5-nitro-2-furoic acid, HMF- 5-hydroxymethylene-2-(5H)-furanone, HFA-5-hydroxy-2-furoic acid.
CH2COOH Cl
H
N
Cl
Dantas et al. [55] investigated the toxicity of the the evolution of toxicity was similar but with a faster
ozonated solution of bezafibrate by using V. fischeri decay from the maximum value.
strains, and expressed it in ‘‘effect concentration’’ Pérez-Estrada et al. [50] observed higher toxicity on V.
(EC50,15min) values (percentage of initial dilution, % v/v, fischeri when treating dipyrone by TiO2-assisted photo-
that causes a 50% reduction of bacteria bioluminescence catalysis than by photo-Fenton photocatalysis, although
in 15 min). After a sharp increase in toxicity in the first the same intermediate products were formed. This was
5–7 min, inhibition of bacteria slowly decayed reaching due to the fact that, in TiO2 photocatalysis, the amount
a value lower than that initially measured for bezafibrate of 4-methylaminoantipiryne (MAA), the hydrolytic
solution after 105 min of ozonation. product of dipyrone, was higher, and, when only MAA
In another study, the same authors [58] monitored was present at t = 0, toxicity was already at the 50%
the abatement and the changes in acute toxicity of threshold. The toxicity determined for photo-Fenton
bacteriostatic, sulfonamide-type antibiotic sulfamethox- treatment was always below the 50% bioluminescent
azole during ozonation. The toxicity profile determined bacteria-inhibition threshold, whereas, in the TiO2-pho-
from the EC50 values calculated showed an increase in tocatalytic treatment, it was found to be higher. Atten-
acute toxicity in the first 30 min of ozonation, reaching tion should therefore be paid when discharging MAA
the initial value of sulfamethoxazole at the end of the into the environment.
experiment (1 h). This indicated formation of interme- Similar to previously mentioned studies, in TiO2-pho-
diates that were more toxic than the original drug. tocatalytic treatment of salbutamol, the highest toxicity
Furthermore, poor mineralization (only 18% of TOC re- observed (i.e. 54%) coincided with the highest amount of
moved) attained after 1 h of ozonation indicated that by- degradation products measured [54]. Considering that
products formed towards the end of the reaction time the toxicity of the initial solution of salbutamol was only
were not susceptible to O3 attack. These end-products 4%, attention should be paid when optimizing the TiO2-
were probably of lower aromatic character (e.g., alde- photocatalytic treatment since more hazardous products
hydes, ketones and organic acids), since the UV254 are formed. Fig. 5 shows the evolution of toxicity along
absorbance removal recorded was >80%. Additionally, the TiO2-photocatalytic degradation of salbutamol.
biodegradability, measured as biological oxygen demand Besides the V. fischeri bioluminescence assay, a short-
(BOD5)/COD ratio was constantly increasing, changing term 4–24 h D. magna feeding-response assay is also
from 0 to 0.28 in the time span of the experiment. applied for determination of acute toxicity. It is based on
Medana et al. [53] also applied a Microtox assay to algae-clearance rates, since feeding in D. magna and
evaluate the toxicity of TiO2-photocatalytic products of many other aquatic organisms is physiologically linked
atenolol. The initial toxicity of atenolol solution showed with growth and reproduction, hence effects on feeding
a negative inhibition value of 15%, due to a biolumi- are usually translated to population responses [61].
nescence-stimulation effect, known as hormesi [59]. Beltrán et al. [62] compared the efficiencies and tox-
During the first 60 min of bioassay, the effect remained icity evolution of the aqueous solution of sulfamethox-
negative, reaching a value around 0% after 4 h.
Méndez-Arriaga et al. [56] investigated the efficiency
of TiO2-photocatalytic treatment in degrading three non-
steroidal anti-inflammatory drugs (NSAIDs) (i.e. ibu-
profen, diclofenac and naproxen) under varying condi-
tions of temperature, TiO2 loading and dissolved-oxygen
concentration. Several hydroxylated derivatives of ibu-
profen were identified as primary degradation products,
to be followed by demethylation and decarboxylation.
The increase in inhibition of bioluminescence was ob-
served in the first 120 min and 240 min in the solution
with excess and lack of oxygen, respectively, due to the
formation of toxic by-products. Interestingly, the authors
reported that the intensity of by-products in the non-
saturated conditions was double that observed for the
oxygen-saturated solution.
Li et al. [60] observed formation of more toxic initial
intermediates of ozonation of oxytetracycline, with an
inhibitory effect on bioluminescence of V. fischeri
increasing up to 99.5% after the first 60 min at pH 3, Figure 5. Inhibition (%) of the luminescence of bacteria Vibrio fisc-
heri as a function of photocatalytic treatment time (reproduced
with a dramatic decrease to 32% in the next 60 min.
from [55] with permission of Elsevier).
Under neutral and basic conditions (i.e. pH 7 and 11),
azole treated by UVA radiation, ozonation, catalytic alterations in immune and endocrine systems, genotoxic
ozonation (O3/TiO2), ozone photolysis (O3/UVA), pho- effects and reproduction-rate decline [67,68]) performed
tocatalytic ozonation (O3/TiO2/UVA) and photocatalytic with secondary wastewater effluent. For example, Petala
oxidation (O2/TiO2/UVA). Acute-toxicity D. magna as- et al. [69] used the Ames test together with the Microtox
says were conducted using the commercial test kit assay to estimate the mutagenicity of secondary waste-
Daphtoxkit F, in accordance with OECD guideline 202 water effluent treated with ozone. The Ames assay is a
(Daphnia sp. Acute Immobilization Test) [63]. The re- bacterial reverse-mutation assay that evaluates the
sults indicated that oxidation of sulfamethoxazole led to genotoxicity of a compound by measuring its ability to
intermediates that still exhibited some sort of toxicity induce reverse mutations at selected loci in a bacterial
[62]. The toxicities observed at the end of the experi- strain. Usually, the compensating mutation must occur
ments were most probably due to unknown intermedi- by the same mutagenic mechanism, so mechanistic
ates, since the effect of the detected concentrations of toxicology information is available from Ames-assay re-
carboxylic acids (oxalic, maleic and fumaric acid) were sults based on the pattern of which strain reverted. It
below the observed EC50 value. was found that ozonation at mild conditions (i.e. low O3
However, these short-term acute-toxicity assays are not doses and short ozonation times) augmented mutage-
always adequate indicators of toxicity. For example, the nicity, possibly due to the formation of intermediate TPs,
Microtox test, developed in the 1980s, is primarily testing whereas, by intensifying the ozonation conditions, these
lethality (like many other currently applied bioassays) (i.e. compounds were further degraded to non-mutagenic
acute toxicity rather than chronic toxicity). With the products.
development of highly sensitive, selective analytical Nakamura et al. [70] investigated mutagenic effects of
equipment measuring ng/L and pg/L concentrations of chlorination products of two nitrofuran antibiotics
contaminants, the assessment of long-term toxicity, (nitrofurazone and frazolidone) by Ames assay [i.e.
potentially causing subtle ecotoxicological effects, seems exposure to Salmonella typhimurium strain TA100 with-
more relevant than predicting acute toxicities at the mg/L out rat-liver homogenate (S9 mix)]. The S. typhimurium
level. The extrapolation of laboratory-based determina- strain used in the assay has a unique mutation that has
tions of lethal concentrations (LCs) to the effects of phar- turned off histidine biosynthesis in the bacteria, which
maceuticals at environmentally-relevant ng/L levels is can then undergo a reverse mutation turning the
uncertain, whereas subtle effects, other than threshold essential gene back on to permit the cell to grow in the
toxicity, would be expected for environmentally-relevant absence of either histidine or tryptophan (prototrophy).
concentrations of pharmaceuticals [64]. More impor- The bacteria use the trace histidine or tryptophan to
tantly, potential modes of action (MOAs) of pharmaceu- undergo several cell divisions, but will stop growing once
ticals on non-target aquatic organisms could differ they have run out, leaving a characteristic ‘‘background
significantly from the MOAs investigated when conduct- lawn’’ that decreases in density with increasing toxicity.
ing tests, such as Microtox and other acute-toxicity tests. After 48 h, only those cells that have undergone a re-
As observed by Sanderson and Thomsen [65], the ap- verse mutation turning the essential gene back on have
proach applied so far by the scientific community can be survived, producing mutant colonies. The background
described by the ‘‘lamppost analogy’’ – ‘‘Looking for the lawn density is scored and the number of revertant
car keys under the street lamppost – even though they got colonies then counted. Mutation results are reported as
lost by the buildingÕs front door in the dark’’. revertants per plate. For the chlorinated solution of
Furthermore, Seiler [66] pointed out that even slight, frazolidone, Nakamura et al. [70] recorded no mutagenic
non-significant influences on single components within response, whereas chlorination by-products of nitrofu-
regulatory cascades (e.g., cellular division or signal razone [i.e. 5-nitro-2-furaldehyde (NFA), 5-nitro-2-
transduction), which would not result in any acute furoic acid (NFOA), 5-hydroxy-2-furoic acid (HFA),
effect, might ultimately adversely affect the entire pop- 5-hydroxymethylene-2-(5H)-furanone (HMF), and chlo-
ulation through sequential propagation or interactions rinated product of nitrofurazone] exhibited mutagenic
with additional, unrelated factors. Acute-toxicity bioas- effects on the TA100 strain without the S9 mix.
says conducted on a few selected eco-organisms could be Short-term, sub-lethal, toxicity assays are frequently
insufficient to predict the hazardousness of a compound, applied to predict long-term ecological effects of toxic
especially in the case of pharmaceuticals, the potential compounds [71]. Palominos et al. [72] investigated the
effects of which may not be detected by such tests. Seiler susceptibility of Escherichia coli strains inoculated at agar
et al. [66] proposed consideration of pharmacodynamic plates to inhibition by reaction products of TiO2-photo-
properties of drugs, necessary effect concentrations and catalytic degradation antibiotic flumequine. A clear
range of activities in order to derive a potential ecotox- correlation between depletion of antibiotic and total
icological profile. antibacterial activity was found, since the inhibition halo
There has been some assessment of adverse effects around the microdrop seeded at the agar plate disap-
related to sub-lethal effects (e.g., mutagenic effects, peared completely for samples after 30 min of irradia-
tion, the time that was necessary for complete elimina- primary liver cancer cells HepG2 for the determination of
tion of flumequine. toxicity of sulfamethoxazole and its ozonation products,
Andreozzi et al. [73] reported ozonation to be an effi- due to their long proliferation time. The ozonation
cient technique for reducing the toxicity of antibacterial products of sulfamethoxazole, including sulfanilamide
agent lincomycin. An Agar diffusion test and liquid (active agent of Prontosil and the first sulfa drug dis-
growth-inhibition tests against lincomycin performed covered), methanol, ethanol and phenol, stimulated
with different microalgal strains revealed EC50 values proliferation of HepG2 and an increase in cell metabo-
varying from 195 lg/L for S. leopoliensis to 1.51 mg/L for lism, instead of deteriorating it. But, it was impossible to
P. subcapitata and 1.63 mg/L for C. meneghiniana, distinguish which degradate was provoking those
whereas the resulting ozonation samples were less toxic changes or which cell functions were affected by the
after only 1 h of treatment. morphological changes observed.
Baran et al. [74] investigated the toxicity of sulfona- However, Bedner et al. [18] reported two toxic deg-
mides (sulfacetamide, sulfamethoxazole, sulfadiazine and radation products of acetaminophen as a result of
sulfathiazole) and their TiO2-photocatalytic degradation treatment with hypochlorite, NAPQI and 1,4-benzoqui-
products in experiments with green alga Chlorella vul- none. NAPQI is a toxic metabolite of acetaminophen
garis. Their growth was monitored by measuring spec- formed in liver at higher administered doses [78], pro-
trophotometrically at 680 nm the optical density of the voking hepatic necrosis. This compound readily hydro-
cell suspension with the added treated sample after 48 h lyzes in aqueous solution to 1,4-benzoquinone [79], a
absorbance, and comparing it to the control sample. It benzene metabolite implicated in genotoxic and muta-
was found that all four sulfonamides were toxic to C. genic effects [80]. As regards the evolution profiles of the
vulgaris, whereas photocatalytic degradation products toxic products of acetaminophen chlorination, NAPQI
were estimated to be less toxic than the original drugs. reached its maximum concentration after 10 min, and
Andreozzi et al. [75] performed toxicity tests by then decayed to form 1,4-benzoquinone. The concen-
exposing algae (S. leopoliensis) and invertebrates tration of 1,4-benzoquinone increased over time to
(Brachionus calyciflorus-rotifer) to aqueous samples con- 25% of the initial concentration of acetaminophen
taining the mixture of six pharmaceuticals (i.e. car- after 1 h of chlorination, while, at that time for NAPQI, it
bamazepine, clofibric acid, diclofenac, sulfamethoxazole, was only 1.5%. In the study where acetaminophen,
ofloxacin and propranolol) treated for different reaction NAPQI and 1,4-benzoquione were intraperitoneally in-
times by oxidative techniques (i.e. ozonation, H2O2/UV jected into mouse, their LD50 toxicities were 500 mg/kg,
and TiO2 photocatalysis). Ozonation and UV/H2O2 re- 20 mg/kg and 8.5 mg/kg, respectively [79]. In other
duced the toxicity on rotifers after 2 min and 3 min, words, NAPQI and 1,4-benzoquinone were 25 times and
respectively, whereas for algae this time was even 58 times more toxic than acetaminophen itself. NAPQI
shorter (1 min). Moreover, after 3–5 min, the growth of and 1,4-benzoquinone are not likely to persist in
algae was stimulated, possibly due to the usage of wastewater treatment where sulfite is employed during
intermediates formed as a carbon source. However, dechlorination, since they would probably be reduced
TiO2-photocatalytic treatment displayed the formation of into acetaminophen and 1,4-hydroquinone, respectively.
toxic intermediates, since, even after 48 h, the inhibition In their study on the UV/H2O2 oxidation of carbam-
effect on algae and invertebrates was observed to be 69– azepine, Vogna et al. [81] reported acridine intermedi-
83% and 75–85%, respectively. ates as the main products of the radical-degradation
Since pharmaceuticals and their degradation products pathway, which are more toxic than the parent com-
may affect organisms from different trophic levels, tox- pound due to their mutagenic and carcinogenic activity
icity should be evaluated by a battery of different bio- [82], so failure to ensure complete mineralization of
assays, since the intermediates formed will exhibit carbamazepine in UV/H2O2 treatment may worsen
different toxicities, depending on the test applied. For environmental and health impacts of water contamina-
example, bioassays with V. fischeri, D. magna and tion by carbamazepine.
Microalga measured different changes in the toxicity of The change in toxicity during processes such as
the solution of pesticides during photo-Fenton photoca- chlorination, ozonation and AOPs will depend on the
talysis and TiO2 photocatalysis, for the same levels of mechanism of reaction of a certain compound (i.e.
TOC [76]. whether the moieties essential to the activity of the
High-throughput assays that use cells or cell lines, molecule react or not). For example, Lange et al. [83]
preferably of human origin, to evaluate the toxicities of reported that ozonation of macrolide antibiotics, namely
compounds may measure relatively simple processes clarithromycin, leading to N-oxidation of the dimethyl-
(e.g., binding of environmental agent with cellular pro- amino group inactivates these drugs, as assayed by
tein and changes in gene expression caused by that suppression of the growth of Pseudomonas putida. The
binding) or more integrated responses (e.g., cell division macrolide antibiotics are thought to block the tunnel
and cell differentiation). Yargeau et al. [77] used human that channels the nascent peptides away from the pep-
Table 2. Antibacterial substrates and expected sites of O3 attack. Adapted from Dodd et al. [87].
Macrolides
O O O
N O
HO
ROX O3 CH2CHO
O3 O
O
HO OH N
O
O
OH HO O3 O
O N O3 OH
O
HO OH
O OH HO
O O
5 O HO
N N O OH
HO
O3
O O C2H5 O
O O O O
OCH3
Tylosin
pKa=7.7
O OH O O
O
OH
Roxithromycin Azithromycin
pKa=9.2 pKa 1,2=8.7, 9.5
DHFRa Inhibitor
O
Sulfonamides NH2
O O O3
O3
O3 H H OCH3
N S NH N S NH
N
O O
N N H2N N OCH3
O O3 O O3
O3
OCH3
Sulfamethoxazole N(4)-acetyl-sulfamethoxazole Trimethoprim

pKa 1,2=1.7, 5.6 pKa=5.5 pKa 1,2=3.2, 7.1
Fluoroquinolones OH
Lincosamide
C2H5
O3 O3
HN N
N N N N
H
N
O3 O3
S O N
OH OH
F F
O O3
O3
O O OH
O O
Ciprofloxacin Enrofloxacin Lincomycin
pKa 1,2=6.2, 8.8 pKa 1,2=6.1, 7.7 pKa=7.8
ß-lactams Tetracycline
O3 O3 O3
HN O
OH
N
S
S OH O3
O N
N
H
O N NH2
NH2
OH O
O3
O
O3 OH O OH OH O
O3
O
HO
O3
Penicilin G Cephalexin Tetracycline

pKa=2.7 pKa 1,2=2.5, 7.1 pKa 1,2,3=3.3, 7.7, 9.7
Aminoglycoside O3
OH OH OH
O O OH
Amikacin O O
pKa 1,2,3,4=6.7, 8.4, 8.4, 9.7 HO HN NH2 OH
HO O NH2
O3
NH2
a
DHFR, dihydrofolate reductase.
tidyl-transferase center inside the ribosome. For this ECOSAR can be downloaded freely from the US EPA
function, the dimethylamino group of the desosamine website (http://www.epa.gov/oppt/exposure/docs/epi-
sugar is considered the major domain necessary for suitedl.htm).
binding the macrolide to its target [84], so N-oxidation of Buth et al. [29] used ECOSAR to determine the pre-
macrolides would deactivate them as ribosomal antibi- dicted no-effect concentration (PNEC) of cimetidine and
otics. Acetalization of the keto group of clarithromycin its chlorination products. From the toxicity predicted by
was also noted during ozonation, which was also con- ECOSAR, intermediates 4-hydroxymethyl-5-methyl-1H-
sidered to lower the biological activity [84]. imidazole and 4-chloro-5-methyl-1H-imidazole were
Dodd and Huang [85] reported that, during chlori- estimated to be more toxic than cimetidine itself,
nation of trimethoprim, halogenation and/or oxidation whereas 4-chloro-5-methyl-1H-imidazole had a PNEC
reactions took place at the 2,4-diaminopyrimidinyl derived from LC50 of 6.5 lg/L, the same order of mag-
moiety, leading to the formation of (multi)chlorinated nitude as for atrazine and 2,4-dichlorophenol [29].
and hydroxylated products. Since the antibacterial However, PNECs calculated for cimetidine sulfoxide and
activity of trimethoprim is derived from its 2,4-diamino- both ß-sultam and d-sultam (based on EC50) were rela-
5-methylpyrimidine moiety, which blocks bacterial fo- tively high (of the order of 100 lg/L). Moreover, the
late synthesis by occupying available dihydrofolate experiments performed with NH4Cl as quenching agent
reductase enzymes [86], and this group is extensively of free chlorine demonstrated that the same TPs are
substituted in reactions with free chlorine, it is likely that likely to be formed even during wastewater treatment.
chlorination will reduce the antibacterial activity. The intermediates formed during drinking-water and
As argued by Dodd et al. [87], some antibiotics may wastewater treatment deserve special attention, even
react in ozonation preferentially with moieties not when their measured toxicities and biological activities
essential to the biochemical activities of the parent are of minor importance. For example, an N-chlorinated
molecule, or they can simply be refractory to O3. For intermediate identified as a product of chlorination of
example, the flumequine moiety in fluoroquinolone sulfamethoxazole was observed to decay by ring chlori-
antibiotics ciprofloxacin and enrofloxacin (see Table 2) nation or back reaction to the parent compound [30].
will react with O3 14 times and 109 times slower, Considering that dechlorination is a common practice in
respectively, at pH 7 than their non-essential piperazine most WWTPs, this could lead to release of sulfameth-
moiety, so, in these cases, loss of an original drug does oxazole from its N-chlorinated product.
not necessarily correspond to loss of antibacterial activ-
ity. Also, essential functional groups of N4-acetyl-sulfa-
methoxazole and ß-lactams are possibly not sufficiently 6. Conclusions and future trends
oxidized during ozonation. For ß-lactam penicillin G,
reaction with the OH radical will be more desirable than Even though disappearance of the original drug is easily
oxidation by O3, since it will result in biochemically- achieved in chlorination, ozonation and AOPs, by-
inactive products benzylpenilloic acid and benzylpeni- products generated can be less biodegradable and/or
cilloic acid. Similarly, O3 appears to react only very more toxic than the parent compound. The complex
slowly with the biochemically essential p-sulfonylaniline nature of photolytic transformations and oxidative/
group in N4-acetyl-sulfamethoxazole. However, amika- reductive reactions taking place in such processes
cin should preferentially be degraded by O3, considering underscores the need to elucidate structures and poten-
that the OH radical attack occurs at sites not associated tially additive, antagonistic and synergetic ecotoxico-
with the antibacterial activity of the molecule. logical effects of intermediate TPs. The environmental
Another approach to evaluating the toxicity of phar- hazard of pharmaceuticals may be enhanced via asso-
maceuticals and their degradation products is through ciation with complex mixtures of other pharmaceuticals,
modeling programs (e.g., ECOSAR). ECOSAR predicts the their human metabolites and degradation products, as
toxicity of compounds to aquatic organisms (e.g., fish, well as other residues present in municipal wastewater.
invertebrates and algae) by using (Quantitative) Struc- Despite difficulties in understanding the effects that
ture Activity Relationships [(Q)SARs]. The program chronic exposure to pharmaceutical residues have on
estimates an acute (short-term) toxicity and, when human and environmental health, existing studies show
available, chronic (long-term or delayed) toxicity. The that their input into the environment must be controlled
assessment program uses a number of log Kow-based and reduced as far as possible. Regulatory and risk
(Q)SARs to estimate the ecotoxicity of organic com- management relating to pharmaceutical residues in the
pounds for several structural classes, often resulting in environment is complex, taking into account their low
ecotoxicity estimates of a variety of endpoints. For acute risks, uncertain bioavailability and MOAs on non-
example, the (Q)SAR model was found to underestimate target organisms, potentially additive effects in the
the acute toxicity (24 h to 21 days) of pharmaceuticals, presence of degradation products, intrinsic toxicity and
relative to the data measured in standardized tests [88]. bacterial resistance. Ultimately, those challenges should
be faced with case-by-case studies. As information on [20] E. Evgenidou, I. Konstantinou, K. Fytianos, T. Albanis, J. Hazard.
exposure to and biological effects of pharmaceuticals Mater., B 137 (2006) 1056.
[21] C. Sirtori, A. Zapata, I. Oller, W. Gernjak, A. Agüera, S. Malato,
increases, estimations of the safety margin between the Water Res. 43 (2009) 661.
no-effect concentration and expected exposure concen- [22] T. Kosjek, E. Heath, Trends Anal. Chem. 27 (2008) 807.
tration in water will be more precise, and hazard [23] T. Kosjek, E. Heath, M. Petrović, D. Barceló, Trends Anal. Chem.
assessment of any particular drug will become more 26 (2007) 1076.
reliable. [24] X. Zhang, F. Wu, X. Wei Wu, P. Chen, N. Deng, J. Hazard. Mat.
157 (2008) 300.
[25] R. Andreozzi, V. Caprio, R. Marotta, D. Vogna, Water Res. 37
(2003) 993.
Acknowledgments [26] D.A. Lambropoulou, M.D. Hernando, I.K. Konstantinou, E.M.
Thurman, I. Ferrer, T.A. Albanis, A.R. Fernández-Alba, J.
The work was financially supported by the Spanish
Chromatogr., A 1183 (2008) 38.
Ministry of Science and Innovation, projects SOSTAQUA [27] L.A. Pérez-Estrada, S. Malato, W. Gernjak, A. Agüera, E.M.
(led by Aguas de Barcelona and founded by CDTI in the Thurman, I. Ferrer, A.R. Fernández-Alba, Environ. Sci. Technol.
framework of the Ingenio 2010 Programme under the 39 (2005) 8300.
CENIT call) and CEMAGUA (CGL2007-64551/HID), and [28] D.C. McDowell, M.M. Huber, M. Wagner, U. Von Gunten, T.A.
Ternes, Enivron. Sci. Technol. 39 (2005) 8014.
the Unity Through Knowledge Fund (UKF), which was
[29] J.M. Buth, W.A. Arnold, K. McNeill, Environ. Sci. Technol. 41
established by the Croatian Ministry of Science, Educa- (2007) 6228.
tion and Sports through the World Bank Loan No. 7320- [30] M.C. Dodd, C.H. Huang, Environ. Sci. Technol. 38 (2004) 5607.
HR. J.R. gratefully acknowledges the JAE Program (Junta [31] T.E. Doll, F.H. Frimmel, Water Res. 38 (2004) 955.
para la Ampliación de los Estudios), co-financed by CSIC [32] T.E. Doll, F.H. Frimmel, Catal. Today 101 (2005) 195.
[33] M. Bedner, W.A. MacCrehan, G.R. Helz, J. Chromatogr. Sci. 40
(Consejo Superior de Investigaciones Cientı́ficas) and
(2002) 447.
European Social Funds, for a predoctoral grant. [34] S. Pérez, P. Eichhorn, M.D. Celiz, D.S. Aga, Anal. Chem. 78 (2006)
1866.
[35] P. Eichhorn, P.L. Ferguson, S. Pérez, D.S. Aga, Anal. Chem. 77
References (2005) 4176.
[1] R.B. Silverman (Editor), The Organic Chemistry of Drug Design [36] G. Hopfgartner, E. Varesio, V. Tschappat, C. Grivet, E. Bourgogne,
and Drug Action, Academic Press, San Diego, CA, USA, 1992. L.A. Leuthold, J. Mass Spectrom. 39 (2004) 845.
[2] J. Radjenović, M. Petrović, D. Barceló, Trends Anal. Chem. 26 [37] T. Paul, P.L. Miller, T.J. Strathmann, Environ. Sci. Technol. 41
(2007) 1132. (2007) 4720.
[3] J.P. Sumpter, Toxicol. Lett. 102–103 (1998) 337. [38] B. De Witte, J. Dewulf, K. Demeestere, V. Van De Vyvere, P. De
[4] K. Kummerer, J. Antimicrob. Chemother. 54 (2004) 311. Wispelaere, H. Van Langenhove, Environ. Sci. Technol. 42 (2008)
[5] B. Halling-Sørensen, S.N. Nielsen, P.F. Lanzky, F. Ingerslev, H.C.H. 4889.
Lützhøft, S.E. Jørgensen, Chemosphere 36 (1998) 357. [39] P. Calza, M. Pazzi, C. Medana, C. Baiocchi, E. Pelizzetti, J. Pharm.
[6] C.G. Daughton, T.A. Ternes, Environ. Health Perspect. 107 Biomed. Anal. 35 (2004) 9.
(1999) 907. [40] P. Calza, V.A. Sakkas, C. Medana, C. Baiocchi, A. Dimou, E.
[7] D.B. Huggett, B.W. Brooks, B. Peterson, C.M. Foran, D. Schlenk, Pelizzetti, T. Albanis, Appl. Catal., B 67 (2006) 197.
Arch. Environ. Contam. Toxicol. 43 (2002) 229. [41] H.G. Wetzstein, N. Schmeer, W. Karl, Appl. Environ. Microbiol. 63
[8] A. Villegas-Navarro, L.E. Rosas, J.L. Reyes, Comp. Biochem. (1997) 4272.
Physiol., C: Toxicol. Pharmacol. Endocrinol 136 (2003) 127. [42] S. Bielecka-Grzela, A. Klimowicz, J. Clin. Pharm. Therap. 30
[9] M. Cleuvers, Chemosphere 59 (2005) 199. (2005) 383.
[10] B.I. Escher, N. Bramaz, M. Richter, J. Lienert, Environ. Sci. [43] D.T.W. Chu, P.B. Fernandes, Antimicrob. Agents Chemother. 33
Technol. 40 (2006) 7402. (1989) 131.
[11] N. Laville, S. Ait-Assa, E. Gomez, C. Casellas, J.M. Porcher, [44] H.J. Zeiler, U. Petersen, W. Gau, H.J. Ploschke, Arzneimittel-
Toxicology 196 (2004) 41. Forschung/Drug Res. 37 (1987) 131.
[12] E.A. Tendencia, L.D. De La Pena, Aquaculture 195 (2001) 193. [45] P. Calza, C. Medana, M. Pazzi, C. Baiocchi, E. Pelizzetti, Appl.
[13] C. Mimeault, A.J. Woodhouse, X.S. Miao, C.D. Metcalfe, T.W. Catal., B 53 (2004) 63.
Moon, V.L. Trudeau, Aquat. Toxicol. 73 (2005) 44. [46] I. Dalmázio, M.O. Almeida, R. Augusti, T.M.A. Alves, J. Am. Soc.
[14] K. Ikehata, N.J. Naghashkar, M.G. El-Din, Ozone Sci. Eng. 28 Mass Spectrom. 18 (2007) 679.
(2006) 353. [47] J.M. Campbell, B.A. Collings, D.J. Douglas, Rapid Comm. Mass
[15] O. Legrini, E. Oliveros, A.M. Braun, Chem. Rev. 93 (1993) 671. Spectrom. 12 (1998) 1463.
[16] T. Oppenlander (Editor), Photochemical Purification of Water and [48] J. Radjenović, C. Sirtori, M. Petrović, D. Barceló, S. Malato,
Air: Advanced Oxidation Processes (AOPs): Principles, Reaction Environ. Sci. Technol. (2009), submitted for publication.
Mechanisms, Reactor Concepts, Wiley-VCH Verlag, Weinheim, [49] J. Radjenović, C. Sirtori, M. Petrović, D. Barceló, S. Malato, Appl.
Germany, 2003. Catal. B. (2009), in press, doi:10.1016/j.apcatb.2009.02.013.
[17] F.J. Beltrán (Editor), Ozone Reaction Kinetics for Water and [50] L.A. Pérez-Estrada, S. Malato, A. Agüera, A.R. Fernández-Alba,
Wastewater Systems, Lewis Publishers, Boca Ratón, FL, USA, Catal. Today 129 (2007) 207.
2003. [51] A.G. Marshall, C.L. Hendrickson, G.S. Jackson, Mass Spectrom.
[18] M. Bedner, W.A. MacCrehan, Environ. Sci. Technol. 40 (2006) Rev. 17 (1998) 1.
516. [52] A. Makarov, E. Denisov, A. Kholomeev, W. Balschun, O. Lange, K.
[19] R. Rosal, A. Rodrı́guez, J.A. Perdigón-Melón, A. Petre, E. Garcóa- Strupat, S. Horning, Anal. Chem. 78 (2006) 2113.
Calvo, M.J. Gómez, A. Agüera, A.R. Fernández-Alba, Chemo- [53] C. Medana, P. Calza, F. Carbone, E. Pelizzeti, H. Hidaka, C.
sphere 74 (2009) 825. Baiocchi, Rapid Comm. Mass Spectrom. 22 (2008) 301.
[54] V.A. Sakkas, P. Calza, C. Medana, A.E. Villioti, C. Baiocchi, E. [72] R. Palominos, J. Freer, M.A. Mondaca, H.D. Mansilla, J. Photo-
Pelizzetti, T. Albanis, Appl. Catal., B 77 (2007) 135. chem. Photobiol., A 193 (2008) 139.
[55] R.F. Dantas, M. Canterino, R. Marotta, C. Sans, S. Esplugas, R. [73] R. Andreozzi, M. Canterino, R.L. Giudice, R. Marotta, G. Pinto, A.
Andreozzi, Water Res. 41 (2007) 2525. Pollio, Water Res. 40 (2006) 630.
[56] F. Méndez-Arriaga, S. Esplugas, J. Giménez, Water Res. 42 (2008) [74] W. Baran, J. Sochacka, W. Wardas, Chemosphere 65 (2006)
585. 1295.
[57] P. Calza, V.A. Sakkas, A. Villioti, C. Massolino, V. Boti, E. [75] R. Andreozzi, L. Campanella, B. Fraysse, J. Garric, A. Gonnella, R.
Pelizzetti, T. Albanis, Appl. Catal., B 84 (2008) 379. Lo Giudice, R. Marotta, G. Pinto, A. Pollio, Water Sci. Technol. 50
[58] R.F. Dantas, S. Contreras, C. Sans, S. Esplugas, J. Hazard. Mater. (2004) 23.
150 (2008) 790. [76] A.R. Fernández-Alba, D. Hernando, A. Agüera, J. Cáceres, S.
[59] R. Grande, S. Di Pietro, E. Di Campli, S. Di Bartolomeo, B. Filareto, Malato, Water Res. 36 (2002) 4255.
L. Cellini, J. Environ. Sci. Health, Part A 42 (2007) 33. [77] V. Yargeau, J.C. Huot, A. Rodayan, L. Rouleau, R. Roy, R.L. Leask,
[60] K. Li, A. Yediler, M. Yang, S. Schulte-Hostede, M.H. Wong, Environ. Toxicol. 23 (2008) 492.
Chemosphere 72 (2008) 473. [78] J.G.M. Bessems, N.P.E. Vermeulen, Crit. Rev. Toxicol. 31 (2001)
[61] C. Barata, D.J. Baird, Aquat. Toxicol. 48 (2000) 195. 55.
[62] F.J. Beltrán, A. Aguinaco, J.F. Garcı́a-Araya, A. Oropesa, Water [79] D.C. Dahlin, S.D. Nelson, J. Med. Chem. 25 (1982) 885.
Res. 42 (2008) 3799. [80] R. Snyder, J. Toxicol Environ. Health. Part A 61 (2000) 339.
[63] Organization for Economic Cooperation and Development (OECD), [81] D. Vogna, R. Marotta, R. Andreozzi, A. Napolitano, M. dÕIschia,
OECD Guidelines for the Testing of Chemicals, Section 2: Effects on Water Res. 54 (2004) 497.
Biotic Systems, Test No. 202: Daphnia sp. Acute Immobilisation [82] S.M. Furst, J.P. Uetrecht, Biochem. Pharmacol. 45 (1993) 1267.
Test, OECD Publishing, Paris, France, 2004. [83] F. Lange, S. Cornelissen, D. Kubac, M.M. Sein, J. von Sonntag, C.B.
[64] E.J. Calabrese, J.W. Staudenmayer, E.J. Stanek, G.R. Hoffmann, Hannich, A. Golloch, H.J. Heipieper, M. Möder, C. von Sonntag,
Toxicol. Sci. 94 (2006) 368. Chemosphere 65 (2006) 17.
[65] H. Sanderson, M. Thomsen, Bull. Environ. Contam. Toxicol. 79 [84] F. Schlünzen, R. Zarivach, J. Harms, A. Bashan, A. Tocilj, R.
(2007) 331. Albrecht, A. Yonath, F. Franceschi, Nature (London) 413 (2001)
[66] J.P. Seiler, Toxicol. Lett. 131 (2002) 105. 814.
[67] S. Aguayo, M.J. Muñoz, A. De La Torre, J. Roset, E. De La Peña, M. [85] M.C. Dodd, C.H. Huang, Water Res. 41 (2007) 647.
Carballo, Sci. Total Environ. 328 (2004) 69. [86] C. Walsh (Editor), Antibiotics: Actions, Origins, Resistance, ASM
[68] H. Dizer, E. Wittekindt, B. Fischer, P.D. Hansen, Chemosphere 46 Press, Washington DC, USA, 2003.
(2002) 225. [87] M.C. Dodd, M.-O. Buffle, U. von Gunten, Environ. Sci. Technol. 40
[69] M. Petala, P. Samaras, A. Zouboulis, A. Kungolos, G.P. Sakellar- (2006) 1969.
opoulos, Water Res. 42 (2008) 4929. [88] H. Sanderson, D.J. Johnson, C.J. Wilson, R.A. Brain, K.R.
[70] H. Nakamura, T. Kawakami, T. Niino, Y. Takahashi, S. Onodera, Solomon, Toxicol. Lett. 144 (2003) 383.
J. Toxicol. Sci. 33 (2008) 621.
[71] C. Barata, P. Alanon, S. Gutierrez-Alonso, M.C. Riva, C. Fernán-
dez, J.V. Tarazona, Sci. Total Environ. 405 (2008) 78.

Complementary Mass Spectrometry and Bioassays

Uploaded by

Copyright:

Available Formats

Complementary Mass Spectrometry and Bioassays

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Complementary Mass Spectrometry and Bioassays

Uploaded by

Copyright:

Available Formats

Trends Trends in Analytical Chemistry, Vol. 28, No.

Complementary mass spectrometry

1. Introduction and diazepam) to no metabolization (e.g.,

WASTE WATER TREATMENT

sewage primary secondary tertiary surface and

ozonation, AOPs chlorination

DRINKING WATER TREATMENT

133.0654 149.0607 [M+H-C(CH3)2-H2O]+ [M+H-H2O]+

116.1066 121.0257 [M+H-C(CH3)2]+

Pharmaceutical, Transformation products Analytical method Toxicity Toxicity test Ref.

Acetaminophen1, NAPQI2, 1,4-benzoquinone3 HPLC/UV/EC, 500 mg/kg1 Intraperitoneal [18]

Sulfamethoxazole N(4)-acetyl-sulfamethoxazole Trimethoprim

Penicilin G Cephalexin Tetracycline

pKa 1,2,3,4=6.7, 8.4, 8.4, 9.7 HO HN NH2 OH

You might also like