Photolysis of Antibiotics under Simulated Sunlight Irradiation

Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES

Article

Photolysis of antibiotics under simulated sunlight irradiation: identification of photoproducts by high resolution mass spectrometry Rosa María Baena-Nogueras, Eduardo Gonzalez-Mazo, and Pablo Antonio Lara-Martín Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.6b03038 • Publication Date (Web): 14 Feb 2017 Downloaded from http://pubs.acs.org on February 20, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 24

Environmental Science & Technology

TOC Art 84x47mm (96 x 96 DPI)

ACS Paragon Plus Environment


1 2 3

Photolysis of antibiotics under simulated sunlight irradiation: identification of photoproducts by high resolution mass spectrometry

4 5

Rosa María Baena-Noguerasa, Eduardo González-Mazoa and Pablo A. Lara-Martín*a

6 7

a

Department of Physical Chemistry, Faculty of Marine and Environmental Sciences, CEI·MAR, University of Cadiz, Puerto Real, 11510, Spain

8 9

* Corresponding author: e-mail: [email protected], phone: +34 956016159, fax: +34 956016040

10 11 12

Abstract

13

There is a growing concern on the widespread use of antibiotics and their presence in

14

the aqueous environment. Their removal in the water column is mediated by different

15

types of degradation processes for which the mechanisms are still unclear. This research

16

is focused on characterizing the photodegradation kinetics and pathways of two largely

17

employed antibiotics families: sulfonamides (9 SDs) and fluoroquinolones (6 FQs).

18

Degradation percentages and rates were measured in pure water exposed to simulated

19

natural sunlight at a constant irradiance value (500 W m-2) during all the experiments,

20

and the main photoproducts formed were characterized through accurate mass

21

measurement using ultra-performance liquid chromatography – quadrupole-time-of-

22

flight – mass spectrometry (UPLC-QToF-MS). Over 100 different phototransformation

23

products were identified for SDs and FQs, 66% of them, to the best of our knowledge,

24

have not been described before. Their sequential formation and disappearance over the

25

course of the experiments reveal the existence of several pathways for the degradation

26

of target antibiotics. Occurrence of new photoproducts derived from desulfonation

27

and/or denitrification as well as hydroxylation of photo-oxidized heterocyclic rings,

28

have been identified during photodegradation of SDs, whereas a new pathway yielding

29

oxidation of the benzene ring after the cleavage of the piperazine ring (e.g. CIP product

30

with m/z 263) is described for FQs.

31 32

Keywords: sulfonamides, fluoroquinolones, photoproducts, pathway, kinetic, high

33

resolution mass spectrometry

1 ACS Paragon Plus Environment

Page 2 of 24

Page 3 of 24

34


1. Introduction

35

The extensive use of antibiotics in human and veterinary medicine is a matter of

36

growing concern because of their role in the emergence and potential spread of

37

resistance genes. More specifically, anthropogenic activities and the inefficient removal

38

of antibiotics and many other pharmaceuticals in wastewater treatment plants (WWTPs)

39

lead to the continuous discharge of these substances into aquatic environments. Among

40

these compounds, fluoroquinolones (FQs) and sulfonamides (SDs) are largely used in

41

human health care systems and in animal farms (1,2), and their increasing detection in

42

WWTP effluents and streams is foreseen as a potential risk for aquatic organisms (3).

43

As examples, levels reported for sulfamethoxazole in San Francisco Bay and

44

Skateneakeles Lake in New York are in the range of 2.4 to 66.7 ng L-1 (4,5), while

45

sulfamehoxazole and ofloxacin occur at levels of up to 169 and 146 ng L-1, respectively,

46

in Ebro River (6,7).

47

Regarding the environmental fate of many antibiotics, including FQs and SDs,

48

degradation processes involving natural light (photodegradation) have been reported as

49

one of the major mechanisms involved in their removal (8-10). Identification of new

50

transformation products of xenobiotics commonly detected in surface waters is of

51

environmental relevance as these chemicals are often ignored in monitoring programs

52

and their toxicity towards aquatic species is unknown. Adverse environmental effects of

53

other pharmaceutical degradation byproducts have been recently highlighted, for

54

example carbamazepine, for which one of the major photoproducts, acridine, is a toxic,

55

mutagenic, and carcinogenic chemical, and therefore, its environmental impact is much

56

higher than that of the parent substrate (11). The photodegradation pathways of most

57

antibiotics, however, are still not fully understood. Several mechanisms have been

58

recently reported to occur during the transformation of specific SDs, such as

59

desulfonation and photohydrolysis for sulfadiazine and sulfapyridine (10,12). Other

60

proposed transformation pathways are the rearrangement of the isoxazole ring and its

61

hydroxylation for sulfamethoxazole (13), as well as the oxidation/reduction of N atoms

62

for sulfadiazine and sulfamethazine (12, 14). Regarding FQs, direct defluorination,

63

hydroxylation of the heterocyclic ring, oxidative degradation of the piperazine moiety,

64

and cyclopropane cleavage at acid pH have been described for ciprofloxacin and

65

enrofloxacin (8,15,16,17,18).



Page 4 of 24

66

So far, most studies have relied on the use of liquid chromatography coupled to

67

triple quadrupole mass spectrometry analyzers (LC-MS/MS) for monitoring the

68

photodegradation of the parent compound and for achieving the identification of

69

possible photodegradation products. This last task is mainly performed through

70

fragmentation of their quasimolecular ions and interpretation of the resulting mass

71

spectra (8,18,19). New generation high resolution mass spectrometers (HRMS) open a

72

new range of possibilities as they allow accurate mass measurement of the

73

quasimolecular ions of possible photoproducts and their specific fragments, enabling

74

unequivocal confirmation of the molecular structures. Therefore, application of

75

advanced HRMS, including a detailed interpretation of the resulting fragmentation

76

pathways at different collision energies, is a powerful tool for relatively rapid and

77

certain structure identification of transformation products, including those from

78

biological reactions (20). New intermediates recently discovered using this approach are

79

desulfonated

80

ciprofloxacin (PT 330) (14,21).

sulfamethazine

(PT

215),

and

defluorinated

and

hydroxylated

81

The aim of this study was to provide a better knowledge on the photodegradation of

82

a selected number of widely used SDs (9) and FQs (6), for many of which there is no

83

available information, in pure water under simulated sunlight irradiation. First,

84

photodegradation kinetics were measured by LC-MS/MS to confirm the disappearance

85

of parent compounds in a short period of time and, later, selected samples were

86

analyzed by ultra-performance liquid chromatography – quadrupole-time-of-flight –

87

mass spectrometry (UPLC-QToF-MS) to elucidate the molecular structure of the new

88

resulting products as well as to confirm the previous proposed photoproducts from

89

studies carried out without using these techniques. Finally, several photodegradation

90

pathways were proposed taking into consideration different photoreaction mechanisms

91

and the sequential appearance/disappearance of several photoproducts over the course

92

of the degradation experiments.

93 94

2. Experimental section

95

2.1. Photolysis experiments

96

Target compounds are listed in Table S1. Photolysis experiments were carried out

97

following the OECD guidelines No. 316 for phototransformation of chemicals in water

98

by direct photolysis. Irradiation was provided by a Suntest CPS+ simulator (Madrid, 3 ACS Paragon Plus Environment

Page 5 of 24


99

Spain) equipped with a xenon lamp which simulates natural sunlight in a wavelength

100

range of 300-800 nm and equipped with coated quartz glass filter. Irradiance was

101

maintained constant at 500 W m-2 during all the experiments and the temperature was

102

monitored, fluctuating between 25 and 35ºC. The solution employed in the first series of

103

experiments consisted of 250 mL of HPLC grade water spiked to 100 ng mL-1 of target

104

compounds. Aliquots of this solution were introduced into 20 quartz tubes of 15 mL

105

each that were placed inside the photo-reactor. Ten of these tubes were covered with

106

aluminum foil and were used as dark controls to account for additional losses from

107

other processes (e.g., hydrolysis and adsorption). Nineteen sampling times were

108

established over a total exposure time of 24 hours, taking two 1 mL aliquots at each

109

sampling time. Once measured, the decreasing concentrations of target compounds

110

versus time were adjusted to the pseudo first-order kinetic.

111

Once photodegradation kinetics were known, a second series of experiments were

112

carried out to identify possible degradation products. These experiments were similar to

113

those aforementioned, but the quartz tubes were filled with individual test solutions of

114

each analyte at 10 µg mL-1. The duration of these experiments was 5 hours, enough to

115

achieve the photodegradation of a significant amount of the parent compounds (from 50

116

to more than 99%), and 13 sampling times were selected. Additionally, pH was

117

monitored in the two series of experiments and remained constant (pH = 6.7) over the

118

course of both assays.

119

2.2. Analysis of samples

120

Simultaneous separation, determination and quantification of parent compounds for

121

the first series of photodegradation experiments were carried out by UPLC-MS/MS

122

using a Bruker Advance - EVOQ system (Bruker Corp., Billerica, MA, USA) following

123

the methodology proposed by Baena-Nogueras et al. (2016) (22). Aqueous samples

124

from the second series of experiments were injected in a UPLC-QToF-MS system, more

125

specifically a Waters Acquity - Synapt G2 tandem (Waters Corp., Milford, MA, USA),

126

for the identification of photodegradation products. Two different functions were use

127

simultaneously in the same run: low energy (4 eV) to obtain quasimolecular ions for

128

parent compounds and photodegradation products, and high energy (energy ramp from

129

10 to 40 eV) to obtain fragments. Further details on sample analysis are described in

130

Supporting Information.

131 4 ACS Paragon Plus Environment


132

2.3. Structure elucidation

133

Due to the relatively high concentrations used in the second series of experiments

134

(10 mg/L), the presence of transformation product peaks could be easily identified by

135

visual inspection against the baseline of the chromatogram. Those having lower

136

abundances were detected through suspect analysis. Identification of these

137

photoproducts was based on the accurate mass measurement (error < 5 ppm) of the

138

detected quasimolecular ions and fragments, as well as the comparison of theoretical

139

and measured isotopic patterns. The MS/MS spectra interpretation was performed using

140

the Mass Fragment tool (MassLynx 4.1 software). Retention times (RTs) of the

141

photoproducts were also considered, being usually lower retention than those for the

142

parent molecules (over 90% of the detected and identified photoproducts had RTs

143

below 3.5 minutes). To clearly communicate the confidence of the proposed

144

transformation products, we assigned confidence levels according to the classification

145

system proposed by Schymanski et al. (2014) (23): identification by exact mass (5),

146

unequivocal molecular formula (4), tentative candidates (3), probable structure (2), and

147

confirmed structure (1). In addition, based on these confidence levels, varying

148

certainties were also attributed to the resulting reactions following the procedure

149

proposed by Gulde et al. (2016) (24): certain, likely, possible and unknown. More

150

detailed information on this can be found in these references and in the columns ‘A’

151

from Tables S2 and S3 (prefix S refers to Supporting Information).

152 153

3. Results and discussion

154

Figures S1 and S2 show the decrease in the concentrations of SDs and FQs over the

155

course of the first series of experiments and their average percentage removals (>98%)

156

at the end of the experiments (24 h), respectively. Table S1 include the kinetic

157

parameters adjusted to the first-order kinetic model which provided a good fit to the

158

experimental data, with coefficients of determination (R2) generally higher than 0.95.

159

Degradation rates (k) and half-lives (t1/2) were in the range of 0.003 to 0.3 min-1 and 2 to

160

227 min, respectively. These two parameters were recalculated (ko and to1/2) considering

161

the effect of other processes (e.g., hydrolysis or sorption onto glass walls) through the

162

analysis of dark controls. This effect was negligible in most cases as differences were

163

below 20% when comparing k and ko (Table S1).


Page 6 of 24

Page 7 of 24


164 165

3.1. Structural elucidation of sulfonamide photoproducts and proposed photodegradation pathways

166

SDs are characterized by the sulfonamide functional group (-RSO2NH2), which

167

consists of a sulfonyl group connected to an amine group. They also have other

168

distinctive features depending on the compound considered, such as heterocyclic ring

169

structures (thiadiazole, isoxazole, and oxazole). Several photodegradation routes have

170

been proposed for the 9 target parent compounds based on the elucidation of the

171

structures of different photoproducts detected by UPLC-Q-ToF-MS, their consecutive

172

formation and elimination, and available information from previous publications. All

173

details on the identification of these photoproducts (n = 102) are shown in Table S2,

174

including retention times, accurate mass measurement, fragments and proposed

175

structures, as well as the mass spectra are displayed in Figure S3. At least 68 out of 102

176

photoproducts identified were described for the first time. More specifically, both direct

177

irradiation and hydrolysis in water led to the degradation of all SDs and the detection of

178

up to 17 degradation products for SMZ, 16 for SMX, 15 for SMP, 13 for SDZ, 10 for

179

STZ, 9 for SGD, SFX and SMT, and 4 for SND. Tentative phototransformation

180

pathways for the 9 SDs under study are presented in Figure 1 and Figures S4 to S11 to

181

illustrate the different mechanisms that take place during the degradation process.

182

Among the first reactions occurring, the extrusions of sulfonyl and amine groups,

183

which can also happen simultaneously, were common for all SDs considered in this

184

study. Boreen and colleagues (2004) determined that photodegradation occurs through

185

SO2 extrusion for SDs containing six-membered heterocyclic substituents, while SDs

186

with five-heterocyclic groups experience a cleavage at various positions (25). The

187

desulfonation process has been also reported (10,12) for some specific compounds such

188

sulfadiazine and sulfapyridine using conventional HPLC, and it was confirmed in our

189

experiments for eight additional SDs. García-Galán and co-workers (2012) (14) also

190

used high resolution mass spectrometry to identify the formation of photoproducts after

191

desulfonation and a combination of desulfonation and denitrification for SMZ and

192

sulfapyridine (m/z 173 and 158). Here, we report for the first time the occurrence of

193

several co-desulfonated and denitrificated products for SMZ (m/z 200, 214, 216), SDZ

194

(m/z 188, 186, 172), SMX (m/z 193, 191, 175), and SMP (m/z 202, 216, 218) (Table

195

S2).



196

Multiple hydroxylated products were also formed during the photolysis

197

experiments, having one or more hydroxyl groups (-OH) that are often positioned in the

198

sulfonamide group (e.g., PT 267 and 283 for SDZ, see Fig. S6) (12,26). Niu and

199

colleagues (2013) (13) have also reported the –OH attack (hydroxylation) on the

200

phenylamine for SMX (PT 272, Fig. 1), which we have confirmed to occur for all the

201

SDs under study: SND (PT 189, Fig. S4), SGD (PT 231, Fig. S5), , SDZ (PT 267, Fig.

202

S6), STZ (PT 272, Fig. S7), SFX (PT 284, Fig. S8), SMT (PT 287, Fig. S9), SMZ (PT

203

295, Fig. S10) and SMP (PT 297, Fig. S11). The hydroxylamine formed may then

204

experience further reduction (27) (e.g., PT 293 for SMZ or PT 265 for SDZ, see Table

205

S2). Additionally, the photo-oxidation of heterocyclic rings has been proposed by

206

Guerard (2009), Niu (2013) (13,27) and co-workers, and observed for some of the

207

compounds tested in this research, such as for the isoxazole ring of SMX (PT 193, Fig.

208

1). New photoproducts derived from all these hydroxylated compounds were identified

209

in our experiments for SMX (PT 270, 256), SMZ (PT 311, 309), SMT (PT 285), SGD

210

(PT 229), STZ (PT 270) and SDZ (PT 265) as a result of successive oxidation and/or

211

reduction reactions. Formation of some of these compounds was also detected in dark

212

controls (e.g., PT 297 from SMP, Table S4), although their signal intensity was often

213

between one and two order of magnitude lower than in irradiated samples, suggesting

214

that light enhances the hydrolysis of SDs. Enhancement of the production of hydrolysis

215

TPs during irradiation has been reported recently for other contaminants such as dioctyl

216

sulfosuccinate (DOSS) (28).

217

Finally, a third photodegradation mechanism involved the cleavage of the

218

sulfonamide group, yielding two major products containing aniline and sulfanilic acids

219

that were previously detected for specific SDs such as SMX, SFX, SMT and STZ (29).

220

Formation of other intermediates such as sulfanilamide (the simplest sulfonamide),

221

aminopyrimidine, or the oxazole ring has been observed for specific SDs depending on

222

their molecular structures. Some SDs were detected as photoproducts of those having a

223

more complex structure and higher molecular weight. Examples are the transformation

224

of SMT into SGD and of SMP into SDZ (Table S2).

225

Many of the reactions described above led to the presence of several

226

chromatographic peaks corresponding to different mono- and di-hydroxylated products

227

(e.g., mono and di-hydroxylated compounds such as the previously identified PT 272

228

and PT 288, respectively, for SMX). Determining the specific position of the hydroxyl 7 ACS Paragon Plus Environment

Page 8 of 24

Page 9 of 24


229

moiety in the molecule was the most challenging part of this study and was achieved by

230

analysis of the fragmentation patterns in the QToF-MS high energy function (see Fig.S3

231

for mass spectra of specific SD intermediates). As an example, high energy spectra

232

allowed us to elucidate the molecular structures of two SMZ photoproducts having the

233

same mass (m/z 295) but observed at different retention times (Table S2).

234

Fragmentation of one of these products resulted in the excision of the hydroxylated

235

aniline (m/z 108), hence the binding of an OH group to this part of the molecule is

236

expected, whereas it led to the loss of just aniline in the other product (Figure 2). The

237

same fragmentation pattern was observed for two other detected photoproducts sharing

238

the same m/z 272 but different retention times in the case of SMX degradation. This

239

approach, however, is limited as fragmentation of a few photoproducts (e.g., the di-

240

hydroxylated SMX product with m/z 288, Table S2) was not observed at the energies

241

selected. Identification in these cases could be improved by using data dependent

242

MS/MS acquisition, although their molecular formulas could be tentatively assigned by

243

accurate mass measurement of their molecular ions and several adducts (e.g., [M+Na]+),

244

an alternative strategy proposed by García-Galán and co-workers (2012) (14). There

245

were other cases (e.g., the identified PT 208, see Fig. 1) where, although fragmentation

246

was feasible and several peaks at different retention times were observed, no distinctive

247

fragments could be identified and therefore the position of the hydroxylated moiety

248

could not be defined.

249

Co-elution of several photoproducts occurred but individual extraction of their

250

respective chromatograms enabled us to outline different evolution curves based on the

251

peak areas (Fig. S12), as we can observe in Figure 3a for SND and its identified

252

photoproducts with m/z 109, 110 and 112. During the first 50 min, two different

253

maximum concentrations could be detected for SND hydroxylated products sharing the

254

same molecular ion m/z 189. This can be explained considering two different oxidation

255

processes such as one over the terminal N atom and another over the benzene ring. The

256

decrease in these photoproducts was followed by an increase in the signal intensities of

257

PT 112 and, later, PT 109 and 110, between 120 and 180 min. After that, the intensities

258

of these compounds became also lower but a further increase could be observed towards

259

the end of the experiment for PT 109 and 110. This could indicate the potential

260

reversibility of some reactions such as dehydroxilation, previously reported by Khaleel

261

and co-workers (2013) (19), and the existence of different degradation pathways 8 ACS Paragon Plus Environment


Page 10 of 24

262

occurring at different speeds (Figure S4). In most cases, and closely related to the

263

oxidation processes aforementioned, there were also photo-reduction reactions resulting

264

in a double bond (from –OH to =O). Equivalent intermediate oxidation steps have been

265

already reported by other authors (10,14) and they were found for all SDs considered in

266

this research.

267

A more detailed description of the evolution in the concentrations of the different

268

photoproducts over the course of the experiments is presented in Figure S12 for SDs,

269

including those intermediates that were also detected in dark controls. This is also

270

summarized in Table S2 (column ‘C’) by using an approach similar to that proposed by

271

Gulde et al. (2016) (24), including a combination of the terms rising (r), steady (s), and

272

falling (f). This strategy also allows assessing the importance of different reactions in

273

terms of how much of the parent compound was transformed via a specific reaction.

274

This is achieved by selecting the highest peak area of each transformation product and

275

relating it to the degree of photodegradation of the parent SD at that same time to

276

calculate the maximal relative amount (column ‘B’ at Table S2). Most SDs were

277

phototransformed through desulfonation reactions showing maximal relative amount

278

values of 91.7, 46, 33.5 and 3.8% for SMP (PT 217), SDZ (PT 187), SMZ (PT 215) and

279

STZ (PT 168). Oxazole ring cleavage of the desulfonated photoproducts (PT 192) was

280

also relevant (10%) for SMX. Other minor reactions that were common for all SDs were

281

oxidation and/or reduction of the heterocyclic ring (e.g., PT 189 from SND) and

282

nitrogen atoms (e.g., PT 229, PT 201, and PT 285 from SGD, SDZ, and SMT,

283

respectively), showing maxima relative amounts up to 2.4%. The rest of reactions were

284

below 1%, except the photohydrolysis of SGD (PT 60) and SMT (PT 116), and the

285

desnitrification of SFX in parallel to hydroxylation (PT 269).

286 287

3.2.

Structural

elucidation

288

photodegradation pathways

of

fluoroquinolone

photoproducts

and

proposed

289

The essential structure of all FQs is formed by a piperazine and a two ring core

290

(containing pyridine and benzene with a fluorine atom at C-7 position) which are

291

responsible for the antimicrobial activity. As in the case of SDs, several

292

photodegradation routes have been proposed for the 6 FQs selected based on the

293

elucidation of the structures of different photoproducts detected by HRMS. All details

294

on the identification of these photoproducts (n = 116) are shown in Table S3, including 9 ACS Paragon Plus Environment

Page 11 of 24


295

retention times, accurate mass measurement, fragments and proposed structures, as well

296

as the mass spectra are displayed in Figure S13 At least 76 out of the 116 photoproducts

297

are described here for the first time. More in detail, the photolysis of FQs led to the

298

detection of up to 26 degradation products for OFX, 24 for CIP, 19 for NOR, 16 for

299

ENR, and SPAR and 15 for DAN. Tentative phototransformation pathways for the 6

300

FQs under study are presented in Figure 4 and Figures S14 to S18 to illustrate the

301

different mechanisms taking place during the degradation.

302

Figure 4 shows the different intermediates that were identified during the

303

photodegradation of ciprofloxacin (CIP), which is one of the most commonly detected

304

antibiotics in wastewater and surface waters (6,30). Up to 24 different photodegradation

305

products were formed as a consequence of direct irradiation and collateral processes

306

such as hydrolysis. Following their changes in intensity over the course of the

307

experiment and their appearance order (Table S5) we were able to propose five different

308

main degradation pathways. The first pathway consists of the hydroxylation (m/z 348)

309

and fluorine solvolysis (m/z 330) of the parent compound, which can take place

310

separately or simultaneously (yielding the detected photoproduct with m/z 346 in the

311

latter case) (21). Wei and coworkers (2013) (15) proposed an alternative structure for

312

PT 346 based on the oxidation of the pyridine ring. The ion formula obtained by

313

accurate mass measurement (C17H20N3O5) as well as the main fragment ions (m/z 328 and

314

300) of this photoproduct matched, however, with the structure proposed by Haddad

315

and Kümmerer (2014) (21). Closely related to PT 346 and PT 348, we could also

316

observe the production of PT 362_2 and 364, derived from a nucleophilic attack over

317

the fluorine atom and two hydroxylations, respectively, although they were detected at

318

lower intensities. The molecular structure for PT 364 was described before (31), not so

319

for PT 362_2.

320

The second route for CIP photodegradation consisted on a direct defluorination

321

(m/z 314) of the parent compound followed by the oxidation of one of the amino groups

322

of the piperazine ring into an amide (m/z 328). Similar products but with two hydrogen

323

atoms less (m/z 312 and 326) were reported before, which may be formed via side-chain

324

oxidation of the piperazine moiety. That would result in intermediates with a double

325

bond previously described not only for CIP but also for enrofloxacin (ENR) (16,18),

326

and also confirmed for the rest of FQs considered in the present research. Previous

327

studies have reported that OH- addition is the most thermodynamically favorable way to 10 ACS Paragon Plus Environment


328

promote the defluorination reaction not only for FQs (15) but also for many

329

haloaromatic substances under irradiation (32). The elucidation of the molecular

330

structure of PT 328 was recently achieved by Haddad and Kümmerer (2014) (21) for

331

ciprofloxacin using a combination of Orbitrap and ion-trap mass spectrometers.

332

Comparison of the mass spectrum obtained in that study with that reported here using

333

QToF-MS reveals a strong fragmentation relationship by the formation of the same

334

fragments, with minor differences (m/z 245 and 231 were recorded only in our case and

335

m/z 213 not) attributed to the use of MS3 in the ion trap in order to gain more structural

336

information.

337

The third degradation route continues both first and second routes described above.

338

It consists of the oxidative degradation of the piperazine ring (see detected

339

photoproducts with m/z 306 and 316_1) (Figure 4). The intermediates formed are

340

derived from two major detected photoproducts, PT 348 (from route 1) and PT 328

341

(from route 2), and, after conversion into PT 306 and 316_1, respectively, end up

342

forming PT 288. This last compound is one of the most abundant CIP products

343

observed, next to PT 330, in agreement with previous studies performed under similar

344

conditions (16,17). Wei and coauthors (2013) (15) also reported the predominance of

345

PT 346, but mostly at acidic and basic pH, whereas PT 306 was only detected at pH 2.

346

In our case, PT 346, together with PT 306 and PT 316_1, were detected over the course

347

of the experiment despite conducting it at neutral pH. Additionally, PT 316_1 yielded

348

PT 344 through a reduction reaction (21).

349

The fourth photodegradation mechanism includes the attack of the piperazine

350

moiety, leading to the occurrence of PT 360, formed by the addition of carbonyl group

351

to the parent compound. This reaction has been previously proposed but the PT 360

352

precursor (PT 362_1) is reported here for the first time (21,31). Both intermediates

353

could also be considered as hydrolysis products as their presence was also detected at

354

trace levels in dark controls (see Figure S19) for the evolution in the concentrations of

355

the different photoproducts over the course of these and other experiments with FQs).

356

CIP can be also photodegraded by cleavage of the piperazine ring, which yields two

357

different fragments (m/z 263, m/z 87). Although this reaction has been observed to take

358

place in darkness, hydrolysis was strongly enhanced by light during the irradiation

359

experiments, as it was previously discussed for SDs, too. Furthermore, PT 263 could

360

undergo oxidation (m/z 279), denitrification coupled to hydroxylation (m/z 280), 11 ACS Paragon Plus Environment

Page 12 of 24

Page 13 of 24


361

fluorine subtraction (m/z 245), and/or reductive dehalogenation (m/z 261) followed by a

362

subsequent nucleophilic substitution of the amino group by a hydroxyl radical (m/z

363

262). The final products (m/z 210 and 194) that were identified have only one ring as

364

the benzene has been totally oxidized. This is the first time that this photodegradation

365

route is described for FQs. Other authors have found different intermediates, such as an

366

ethyl derivative from PT 263, formed by UV radiation (18). There are also some

367

discrepancies between the results reported here and those from previous experiments

368

performed using low resolution mass spectrometry. For instance, the molecular

369

structures reported for PT 210 and PT 194 reported by Babic et al. (2013) (8) were

370

different to those observed here. In that sense, Batchu and collaborators (2014) (9)

371

recently revised the CIP phototransformation pathways and, using HRMS as well,

372

proposed alternative structures for PT 288 and PT 245 consisting of the opening of the

373

pyridine instead of the piperazine ring. Accurate mass measurement of these

374

photoproducts and their main fragment ions (m/z 270 and 227) in our experiments,

375

however, provides a better match with the molecular structures described by Wei et al.

376

(2013) (15). This is also in agreement with the degradation of the piperazine moiety into

377

PT 306 and PT 263, identified before as one of the major routes for the photolysis of

378

H4CIP3+ at pH 2. It seems that in our case, where pH is neutral and zwitterion is the

379

expected dominant form in water, the same mechanism also applies. Additionally, the

380

structure proposed here for PT 263 has been already described during the

381

photodegradation of other FQs such as ENR or DAN (8,,33,34).

382

The fifth alternative route for the degradation of CIP includes its transformation

383

into norfloxacin (NOR) (m/z 320), another FQ, through the cyclopropane cleavage. This

384

type of mechanism has been previously observed for other FQs as they share very

385

similar structures (8), such as the transformation of enrofloxacin (ENR) into CIP

386

(Figure S14). More specifically, photolysis experiments carried out with ENR resulted

387

in the identification of 16 irradiation products plus those derived from CIP (Figure 4).

388

The occurrence of CIP during ENR degradation, however, could not be detected during

389

tests performed at acidic pH (8). In our case, ENR followed several degradation

390

pathways that were very similar to those described for CIP, with some new reactions

391

such as the hydrogen abstraction by hydroxyl radical (35) from which PT 376 is

392

transformed into PT 374_2. The photolysis of enrofloxacin has been recently

393

investigated by other groups yielding similar results (18,34) to those presented here, 12 ACS Paragon Plus Environment


394

although some of the photoproducts were different depending on the pH. Examples are

395

the cyclopropane cleavage reported at pH 4 and the oxidative photodegradation at pH 8

396

(8,36). To the best of our knowledge, at least seven of the intermediates described in

397

Figure S11 (PT 378, 374_1, 362, 360, 350, 348, and 344) were not reported previously.

398

More details on the photodegradation of ENR are shown in Figure 3b, describing the

399

evolution in the relative abundance of some of the main products over the course of the

400

experiment and representative chromatograms. Hydroxylated compounds, such as those

401

where fluorine atom was replaced by –OH, have lower retention times (

Photolysis of Antibiotics under Simulated Sunlight Irradiation

Recommend Documents