Chemosphere 93 (2013) 152–159
Contents lists available at SciVerse ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Persistence and degradation of new b-lactam antibiotics in the soil and water environment I. Braschi a,⇑, S. Blasioli a, C. Fellet a, R. Lorenzini a, A. Garelli b, M. Pori b, D. Giacomini b a b
Department of Agricultural Science, University of Bologna, Viale G. Fanin 44, 40127 Bologna, Italy Department of Chemistry ‘‘G. Ciamician’’, University of Bologna, Via Selmi 2, 40126 Bologna, Italy
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Hydrolysis of two new monocyclic b-
lactam antibiotics in the pH range 3– 9. Identiﬁcation of hydrolysis products and hydrolytic pathway. Persistence in a calcareous soil at wilting point and ﬁeld capacity. Persistence in a forest acidic soil at wilting point and ﬁeld capacity. Degradation pathways in soils.
a r t i c l e
i n f o
Article history: Received 20 August 2012 Received in revised form 8 May 2013 Accepted 11 May 2013 Available online 15 June 2013 Keywords: Azetidinones Amoxicillin Hydrolysis Environmental fate Antimicrobial agents
a b s t r a c t The development of new antibiotics with low environmental persistence is of utmost importance in contrasting phenomena of antibiotic resistance. In this study, the persistence of two newly synthesized monocyclic b-lactam antibiotics: (2R)-1-(methylthio)-4-oxoazetidin-2-yl acetate, P1, and (2R,3R)-3((1R)-1-(tert-butyldimethylsilanyloxy)ethyl)-1-(methylthio)-4-oxoazetidin-2-yl acetate, P2, has been investigated in water in the pH range 3–9 and in two (calcareous and forest) soils, then compared to amoxicillin, a b-lactam antibiotic used in human and veterinary medicine. P1 and P2 persistence in water was lower than that of amoxicillin with only a few exceptions. P1 hydrolysis was catalyzed at an acidic pH whereas P2 hydrolysis takes place at both acidic and alkaline pH values. P1 persistence in soils depended mainly on their water potential (t1/2: 35.0–70.7 d at wilting point; <1 d at ﬁeld capacity) whereas for P2 it was shorter and unaffected by soil water content (t1/2 0.13–2.5 d). Several degradation products were detected in soils at both water potentials, deriving partly from hydrolytic pathways and partly from microbial transformation. The higher Log Kow value for P2 compared with P1 seemingly confers P2 with high permeability to microbial membranes regardless of soil water content. P1 and P2 persistence in soils at wilting point was shorter than that of amoxicillin, whereas it had the same extent at ﬁeld capacity. Ó 2013 Elsevier Ltd. All rights reserved.
⇑ Corresponding author. Tel.: +39 051 2096208; fax:+39 051 2096203. E-mail address: [email protected]
(I. Braschi). 0045-6535/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.chemosphere.2013.05.016
I. Braschi et al. / Chemosphere 93 (2013) 152–159
1. Introduction b-Lactam antibiotics such as penicillins and cephalosporins are still the main class of agents used in treating bacterial infections (ESAC, 2009-2010). They share a common mode of action, inhibiting the synthesis of the bacterial cell wall by covalently binding with serine residues in the nucleophilic active site of D,D-transpeptidases (called penicillin-binding proteins, PBPs) (Walsh, 2003). Notwithstanding the pivotal role of b-lactams as ﬁrst choice antibiotics, the treatment of bacterial infections is increasingly complicated as microorganisms can develop resistance to such agents. The emergence of multidrug resistant microorganisms is one of the greatest challenges to the clinical management of infectious diseases. New antimicrobial agents are therefore urgently required. Beside bicyclic b-lactam classes such as penicillins, cephalosporins and carbapenems, monocyclic compounds have emerged due to their interesting, variegated biological activities (Galletti and Giacomini, 2011). The discovery of monobactams has demonstrated that a conformationally constrained bicyclic structure is not essential for antibacterial activity (Walsh, 2000). 4-Alkylidencarbonyl-azetidinones have been discovered to be interesting scaffolds for antibiotics against resistant bacteria (Broccolo et al., 2006), and as effective enzymatic inhibitors against human leukocyte elastase and matrix metalloproteases (Cainelli et al., 2003; Cainelli et al., 2005; Dell’Aica et al., 2006). The efﬁcacy evaluation of some monocyclic b-lactams with an alkylthio group on the b-lactam nitrogen atom, has recently been reported (Turos et al., 2002). More recently still, some monocyclic N-thiomethyl-azetidinones have proved to be interesting as selective inhibitors of histone deacetylases HDAC6 and HDAC8 (Galletti et al., 2009). The potential human and environmental health risks posed by residual antibiotics and their metabolites in surface and groundwater have been receiving growing attention (Kümmerer, 2009a,b). The use of large amounts of non biodegradable antibiotics causes their persistence in water bodies and soils, where they can exert selective pressure for long periods of time (Bergheim et al., 2010). Most resistance determinants have been discovered in clinical and veterinary bacterial isolates but other environmental compartments (i.e. uncultured soils and sewage sludge) have been receiving more attention as reservoirs of new antibiotic resistance genes (Riesenfeld et al., 2004).
In this work, we report on an evaluation of the environmental persistence of two newly synthesized b-lactam antimicrobial agents in the monocyclic azetidinones family: (2R)-1-(methylthio)-4-oxoazetidin-2-yl acetate – P1 – and (2R,3R)-3-((1R)-1(tert-butyldimethylsilanyloxy)ethyl)-1-(methylthio)-4-oxoazetidin2-yl acetate – P2, whose structures and synthesis are reported in Fig. 1. P1 was moderately active against methicillin sensible and methicillin resistant Staphylococcus aureus strains, whereas P2 showed signiﬁcant activity against these strains (Galletti et al., 2011). The P1 and P2 persistence and degradation pathway has been evaluated both in water at different pH values and in two different soils kept at two water potentials. The fate of P1 and P2 has been compared to that of amoxicillin, a broad spectrum b-lactam antibiotic, which has been on the market since the 1970s and is widely used in human and veterinary medicine. 2. Materials and methods 2.1. Monocyclic -lactams Monocyclic b-lactams P1 and P2 were prepared according to procedures previously reported (Galletti et al., 2011) and summarized in Fig. 1. Stock solutions of P1 or P2 (99% and 95% purity, respectively), were prepared dissolving known amounts of each drug (both with a dense, oily appearance) in 5 mL ethyl acetate (RPE ACS grade, Carlo Erba). The amoxicillin trihydrate powder (97% purity) was supplied by Gellini International (Italy). A stock solution at the maximal solubility (1.3 mM) was prepared by dissolving amoxicillin in methanol (J.T. Baker, Holland). 2.2. Soils Two soils from northern Italy (Bologna area) were selected for this study: S1 and S2. After collection of the ﬁrst 10 cm of soils, samples were air dried and the sieved fraction <2 mm was stored at room temperature (RT) in the dark. Soil pH was determined in water or 0.01 M CaCl2 aqueous solution with a soil:solution ratio of 1:2.5 w/w (Conyers and Davey, 1988). CaCO3 content was evaluated using a Dietrich-Freuling calcimeter (Loeppert and Suarez, 1996). Particle size distribution analysis was performed using the hydrometer method (Sheldrick and Wang, 1993). Soil water content at wilting point and ﬁeld capacity (15 and 0.33 bar of suction pressure, respectively) were determined using the gravimetric method at 105 °C for 24 h (Black, 1965). Total organic carbon content (TOC) was determined by CHN analyzer (Model EA 1110 CHNS-O, CE Instruments, UK). Experiments conducted in soils under reduced microbial activity were performed by autoclaving as reported in Alef (1995) and Trevors (1996) treating small soil portions: 50 g of soil were spread in a 1–2 cm layer inside a glass ﬂask, then autoclaved at 121 °C and 1.1 atm for 1 h. 2.3. Hydrolysis kinetics at different pH values
Fig. 1. Synthesis pathway and structures of monocyclic azetidinones P1 and P2. Calculated Log Kow (CLog Kow) values were obtained with ChemDraw 11.0 program (speciﬁc algorithms for calculating Log Kow from fragment based methods were developed by the Medicinal Chemistry Project of CambridgeSoft and BioByte).
The experiments were conducted separately for each antibiotic. Antibiotic solutions were prepared by evaporating known volumes of each antibiotic stock solution to dryness and redissolving them in buffered media at pH 3 or 5 (20 mM trisodium citrate, Carlo Erba), and pH 7 or 9 (20 mM, dibasic potassium phosphate, Fluka). Finally, sodium azide (up to 0.1%, Carlo Erba) was added to maintain sterility). Buffered solutions, at antibiotic concentrations of 100 lM for P1 or 30 lM for P2, were kept in the dark at 25 ± 2 °C on a horizontal shaker for the entire trial duration. The low concentration of P2 was due to its low solubility, which was
I. Braschi et al. / Chemosphere 93 (2013) 152–159
Fig. 2. Hydrolytic degradation pathway of P1 and P2 at pH 9 and 3, respectively, proposed on the basis of LC–MS analysis of by-products. TBS: tert-butyldimethylsilyl.
Table 1 Kinetics parameters (kobs = ﬁrst order rate degradation constant; t1/2 = half life time) for P1 and P2 hydrolysis in the pH range 3–9 at RT. Parameters for amoxicillin are reported for comparison. Mean of three replicates; SD in parenthesis. pH
P1 3 5 7 9
0.0140 0.0506 0.0845 0.7982
(0.0241) (0.0002) (0.0313) (0.0001)
49.5 (0.1) 13.7 (0.1) 8.8 (1.2) 0.87 (0.01)
0.9840 0.9820 0.9820 0.9986
(0.0081) (0.0034) (0.0121) (0.0012)
P2 3 5 7 9
0.4567 0.0270 0.0207 0.2950
(0.0143) (0.0017) (0.0016) (0.0076)
1.6 (0.5) 25.7 (1.5) 33.6 (2.5) 2.4 (0.1)
0.9720 0.9819 0.9580 0.9518
(0.0267) (0.0239) (0.0321) (0.0323)
Amoxicillin 3 5 7 9
0.1106 0.0151 0.0353 0.0481
(0.0094) (0.0018) (0.0023) (0.0028)
6.2 (0.5) 46.1 (5.4) 19.7 (1.3) 14.4 (0.9)
0.9757 0.9700 0.9403 0.9263
(0.0091) (0.0222) (0.0332) (0.0650)
about 35 lM at RT as determined by HPLC. At speciﬁc times (after 2, 4, 6, 8 h for P1 at pH 9; after 2, 4, 6, 8, 24, 26, 28, 30 h for P2 at pH 3; daily for P1 at pH 7 and for P2 at pH 9; after 4, 8, and 12 d for P1 at pH 5; weekly for P1 at pH 3 and for P2 at pH 5 and 7), aliquots of antibiotic solutions were withdrawn and directly analyzed in HPLC. The experiments were run in triplicate. Buffered solutions of amoxicillin (100 lM) at pH 3, 5, 7, and 9 were prepared for comparison. 2.4. Degradation kinetics in soil The experiments were conducted separately for each antibiotic. Experiments on antibiotic degradation in soils S1 and S2 (both untreated and autoclaved) at RT were performed at wilting point and
at ﬁeld capacity. Experiments with the soil alone were also run as a control. All experiments were run in triplicate.
2.4.1. Wilting point Experiments with soil at wilting point were set preparing a number of trials where 5 lL of P1 or P2 stock solutions were placed in polyallomer centrifuge tubes (NalgeneÒ) and, after removal of the solvent by evaporation, added with 2.5 g of S1 or S2, both at wilting point (2.3% and 1.5% soil dry weight – DW – vide infra Table 2). Solvent removal allowed the microbial biomass modiﬁcation induced by solvents in the soils to be limited. The ﬁnal concentrations for P1 and P2 were 0.5 and 0.14 mmol kg1 soil DW, respectively. Attempts at transfering the antibiotics to the soil by homogeneously spraying the solution on each soil top failed in that: (i) the antibiotic partly reached areas of the test tube wall which were not in contact with the soil, and (ii) the solvent never completely desorbed from the soils for the entire duration of the experiment. To maintain an adequate oxygen level for the microbial biomass in the untreated soils, each test tube was sealed with a top which had previously been pierced. Soil humidity was maintained by adding distilled water when necessary in order to restore the water weight lost by evaporation (in general, a few lL of distilled water per week). In the case of soils at reduced microbial activity, each test tube was kept sealed before analysis. To observe soil degradation, suitable MilliQ water volumes were added to centrifuge tubes at speciﬁc time intervals (see Fig. 3 for sampling times), in order to reach the ﬁnal value of 5 mL and kept on a horizontal shaker for 15 min. The supernatants were then separated from the soil particles by two centrifugation cycles at 15 000g for 15 min each and directly analyzed by HPLC. To desorb any chemicals which had possibly been adsorbed onto the soil solid phases, each pellet was shaken for 15 min with a mix-
Table 2 Summary of physicochemical properties of investigated soils (standard deviation – SD – reported in parenthesis). Soil
CaCO3 (g kg1)
55 (6) n.d.
TOC: total organic carbon; DW: dry weight; n.d.: not determined.
TOC (g kg1)
Texture (%) Sand
7.7 (0.7) 21.8 (1.1)
Soil water potential (% soil DW) Wilting point
2.3 (0.2) 1.5 (0.2)
28.0 (2.0) 36.6 (3.4)
I. Braschi et al. / Chemosphere 93 (2013) 152–159
Fig. 3. P1, P2, and amoxicillin degradation in S1 and S2 soils at wilting point and ﬁeld capacity, both untreated and autoclaved, kept at RT. Means of three replicates, SD in vertical bars. When not visible, SD bar is within the symbol dimension. Amoxicillin degradation at wilting point is not reported as no concentration decrease was observed during 3 weeks observation.
ture of acetonitrile and methanol 50%:50% v:v to reach a ﬁnal volume of 5 mL. For each test tube, the amount of residual water after centrifugation was obtained by weighing the tube and the amount of antibiotics contained was then subtracted from the measured concentration. Extraction recovery attested at 97% for P1, and 94% for P2. Each trial was used to give one concentration point: after analysis, the extracted soil and supernatant were discarded. The same trials were conducted in soils under reduced microbial activity. The persistence of amoxicillin in untreated and autoclaved soils at a concentration of 0.2 mmol kg1 soil DW was also performed at
wilting point as a comparison test. Also in this case, due to the low solubility of amoxicillin in water, 835 lL of reference stock solution was placed in the centrifuge test tubes and, after methanol removal by evaporation, 2.5 g of each soil at wilting point was added.
2.4.2. Field capacity Experiments at ﬁeld capacity were set by placing 5 lL of P1 or P2 stock solutions in the centrifuge tubes and, after removal of the solvent by evaporation, adding 2.5 g of S1 and S2 both at ﬁeld capacity (28.0% and 36.6% soil DW – vide infra Table 2). The ﬁnal
I. Braschi et al. / Chemosphere 93 (2013) 152–159
P1 and P2 concentration was 0.5 and 0.14 mmol kg1 soil DW, respectively. The oxygen level for the microbial biomass in the untreated soils was adequately maintained as has already been described in Section 2.4.1. In the case of soils at reduced microbial activity, each test tube was kept sealed before analysis. At speciﬁc times (see Fig. 3 for sampling times), soils were extracted with suitable MilliQ water volumes being added to the centrifuge tubes in order to reach the ﬁnal value of 5 mL and kept on a horizontal shaker for 15 min. Thereafter, soils were extracted and processed in the same way as has been described for the wilting point conditions. The same experiments were conducted in soils under reduced microbial activity. The persistence of amoxicillin in untreated and autoclaved soils at ﬁeld capacity at a concentration of 0.2 mmol kg1 soil DW was performed as a comparison. 2.5. Chromatographic analysis 2.5.1. HPLC–DAD P1 and P2 concentrations were determined by HPLC assembled with a Jasco 880-PU Intelligent pump, a Jasco AS-2055 plus Intelligent sampler, a Jasco 875-UV Intelligent UV–visible detector at 195 nm, a Jones Chromatography model 7971 column heater set at 35 °C, and Borwin Chromatography Software. For P1 analysis, a Synergi 4 u Hydro – RP 80A (150 4.60 mm) eluted with 10% acetonitrile (HPLC grade, J.T. Baker) and 90% MilliQ water at 1 mL min1 was used as an analytical column. Under these chromatographic conditions, the retention times for P1 and its degradation products, named P1-1 (k239nm), P1-2 (k255nm), P13 (k262nm), P1-4 (k327nm), P1-5 (k243nm), P1-6 (k287nm), P1-7 (k277nm) and P1-8 (k283nm) were 12.9, 3.2, 1.7, 2.6, 2.1, 7.3, 5.1, 6.9, and 4.8 min, respectively. P2 was monitored with a Kinetex 2.6u PFP 100A (100 4.60 mm) analytical column eluted with 55% acetonitrile and 45% MilliQ water at 0.3 mL min1, its retention time was 7.20 min. P2 degradation products were observed by applying the P1 analytical set up. Under these chromatographic conditions, the retention times of three P2 degradation products, named P2–1 (k<195nm), P2–2 (k275nm), and P2-3 (k215nm), were 8.1, 5.2, and 3.5 min, respectively. The P1, P2, and P2-1 UV spectra was characterized by intense absorption at wavelengths which were lower than 195 nm and weak absorption in the 237–241 nm region. Though maximal absorption was not centred at 195 nm, the HPLC UV detector was set at this value for the three compounds due to low signal intensity in the 237–241 nm region. Amoxicillin was quantiﬁed at 274 nm with a Waters Spherisorb 5 lm C8 (250 4.6 mm) analytical column eluted with 90% MilliQ water (pH 5 by formic acid, Riedel-de Haën) and 10% methanol (HPLC grade, J.T. Baker) at 1 mL min1 with a retention time of 4.7 min. P1, P2, and amoxicillin quantitative determination was performed by measuring the average value of triplicate injections and by using external standards (rworkingline = 0.999, 0.994, 0.999 for P1, P2, and amoxicillin, respectively). Calculations were based on the average peak areas of the external standards. 2.5.2. LC–ESI–MS single quadrupole The LC–MS equipment was assembled as follows: HPLC Agilent Technologies 1100 series, DAD detector at 210 and 254 nm, Phenomenex Gemini C18 column – 150 4.6 mm 3 i.d.; elution conditions: water:acetonitrile from 70:30 to 20:80 in 8 min at 0.4 mL min1; Agilent Technologies MS single quadrupole 1100 series mod. G 1946B, drying gas ﬂow: 10.5 L min1, nebulizer pressure 25 psgi, drying gas temperature 350 °C, capillary voltage 4500 positive scan, scan range: 100–2600 m/z. The LC–ESI–MS analysis
performed in the positive scan gave the following m/z (%): 133 (100) [MH2O]+ for P1-2 and 461 (50) [2M+Na]+, 242 (100) [M+Na]+, 220 (15) [M+H]+ for P2-1. 2.5.3. LC–ESI–MS ion trap The LC–MS equipment was assembled as follows: Agilent Technologies 1100 series HPLC, DAD detector (UV wavelength 210 and 254 nm), Agilent Technologies Eclipse XDB-C18 column – 150 4.6 mm 3 i.d., elution conditions: water:acetonitrile from 70:30 to 20:80 in 8 min at 0.4 mL min1; Agilent Technologies MS Ion Trap 1100 series mod. G 2445D, drying gas ﬂow: 11.0 L min1, nebulizer pressure: 25 psgi, drying gas temperature 300 °C, capillary voltage 4500 in positive scan and 4000 in negative scan, scan 50-2200 uma. LC–ESI–MS ion trap analysis performed in the negative scan gave the following m/z (%): 246 (100) [M+2H2O+OH] for P1-1 and 182 (100) [M+H2O+OH] for P1-3, 311 (90) [2M+H2O+NaCO3], 293 (100) [2M+NaCO3]- for P1-4. 2.5.4. Data analysis Hydrolysis kinetics data were processed according to the ﬁrst order rate law, ln (C/C0) = kobs t, where C is antibiotic concentration (lM units) at t time, C0 is initial antibiotic concentration, and kobs is ﬁrst order rate degradation constant. It should be noted that the ﬁrst order rate degradation constants related to antibiotic degradation in soils here reported are ‘‘apparent constants’’ (K 0obs ), whose meaning is linked to the speciﬁc situation under consideration. As these constants describe heterogeneous systems (i.e. antibiotics are not homogeneously distributed in the soil samples), they cannot be used in any predictive mathematical model, as they have no physical meaning. Nevertheless, they are descriptive of the evolution of point source contamination caused by antibiotics in the investigated soils. 3. Results and discussion 3.1. Hydrolysis kinetics The P1 and P2 hydrolysis kinetic parameters in buffered aqueous solution at pH values in the 3–9 range at RT are listed in Table 1. As expected, the disappearance of the antibiotics followed a pseudo ﬁrst order kinetics at all investigated pH values (R2 P 0.9263). The P1 degradation rate in water as a function of pH was in the order: 9 > 7 P 5 > 3, with corresponding half life times (t1/2) of 0.87, 8.8, 13.7, and 49.5 d, respectively. P1 hydrolysis resulted thus catalyzed under alkaline conditions, whereas acidity did not seem to have appreciable effects. P2 hydrolysis sped up to such extreme pH values as 3 and 9 (t1/2 1.6 and 2.4 d, respectively), indicating both an acidic and alkaline catalyzed mechanism. Degradation at pH 5 and 7 was slower (t1/2 25.7 and 33.6 d, respectively). With the exception of pH 3, P1 hydrolysis was in general faster with respect to P2. Given that the O(tert-butyldimethylsilyl)-hydroxyethyl side chain attached to the b-lactam ring, which was the only difference between the structures under consideration, the different hydrolytic fate of the two antibiotics can be ascribed to this group, as it is able to increase the hydrophobicity of P2 compared to P1 by four orders of magnitude (CLog Kow 3.8789 and 0.9845, respectively; Fig. 1). The persistence of P1 and P2 in water has been compared to that of amoxicillin, whose widespread use is allowed since it has passed mandatory ecotoxicological tests. As shown by the data reported in Table 1, the persistence of amoxicillin in water as a function of pH was in the order: 5 > 7 > 9 > 3 (t1/2 46.1, 19.7, 14.4, and 6.2 d), thus highlighting weak catalytic activity at both acidic and alkaline conditions. These amoxicillin hydrolysis pathways have
I. Braschi et al. / Chemosphere 93 (2013) 152–159
been known for a great deal of time (Hou and Poole, 1971). It is generally acknowledged that its disappearance is faster in basic solutions (e.g. t1/2 1.2 d at pH 9 and 30 °C, Chada et al., 2003) compared to acidic ones (t1/2 4.7 d at pH 3 and 30 °C, Chada et al., 2003), due to the susceptibility of the b-lactam ring to nucleophilic attack by hydroxide ions (Hou and Poole, 1971). Even if our data are higher (t1/2 14.4 and 6.2 d at pH 9 and 3, respectively), this is most likely due to the different conditions adopted (buffers and temperature adopted, addition of sodium azide to preserve the solutions sterility), as the amoxicillin hydrolysis trend as a function of pH is in agreement with data reported in other studies (Hou and Poole, 1971; Chada et al., 2003). As far as concerns the antibiotics selected for our study, the hydrolytic degradation results can be safely cross compared thanks to the identical experimental conditions adopted. Amoxicillin was more persistent than P1 in aqueous solutions at all the pH values under investigation with the exception of pH 3, where its stability (t1/2 6.2 d) was one order of magnitude lower than P1 (t1/2 49.5 d). Nevertheless natural waters with such acidic pH values are very unusual. On the contrary, waters at pH 7 are more common and, under these conditions, P1 is degraded faster (t1/2 8.8 d) than amoxicillin (t1/2 19.7 d). Amoxicillin was also found to be more persistent when compared to P2 at all investigated pHs except at neutral pH where its t1/2 was calculated at about two thirds of what was observed for P2 (19.7 and 33.6 d, respectively).
3.2. Hydrolysis pathways Hydrolysis product structures (and also soil degradation products, vide infra) were not isolated from the reaction mixture. Rather their structures were hypothesed on the basis of the mass spectra of chromatograﬁcally separated peaks. Given that their molar extintion coefﬁcients were not deﬁned, their concentrations in aqueous (and soil, vide infra) solutions were not determined. However, the kinetics of their formation was monitored by their HPLC peak areas. The formation of P1 hydrolysis products at pH 3 was slow and did not lead, in reasonable times, to appreciable amounts of them. On the contrary, hydrolysis at pH 9 generated four by-products, named P1-1, P1-2, P1-3, and P1-4, which were also present to a lesser extent at pH 7 and 5. P1 hydrolysis product structures are reported in Fig. 2: 3-acetoxy-3-((methylthio)amino)propanoic acid (P1-1), 3-hydroxy-3-((methylthio)amino)propanoic acid (P1-2), 3acetoxy-3-aminopropanoic acid (P1-3), and 3-amino-3-hydroxypropanoic acid (P1-4). According to the degradation pathway, P1 hydrolysis initiates with the opening of the b-lactam cycle with the formation of P11, which in turn loses methylsulﬁde group to give P1-3 or, alternatively, acetic acid to give P1-2. Finally, both P1-2 and P1-3 can be transformed into P1-4 by the loss of an (?) acetic acid or the methylsulﬁde group, respectively. Unfortunately, P2 hydrolysis did not give any appreciable amount of UV visible by-products throughout the investigated pH range with the exception of a single degradation product ((2R,3R)-3-((1R)-1-hydroxyethyl)-1-(methylthio)-4-oxoazetidin-2yl acetate, P2-1), whose formation was observed at pH 3. Its structure is reported in Fig. 2. P2-1 is formed by hydrolysis of the silyloxyl group with the loss of tert-butyldimethylsilanol (TBS-OH) as expected since this group is easily hydrolysable under acidic conditions (Wuts and Greene, 2007). Due to the high reactivity of P2 under alkaline conditions (t1/2 2.4 d, pH 9), the fast degradation of hydrolysis products cannot be excluded either, along with the formation of UV undetectable by-products, as mass analysis of P2 hydrolysis solutions at pH 9 was unsuccessful.
3.3. Persistence in soils at different water retention levels A summary of physical and chemical characteristics of the soils selected for this study is reported in Table 2. S1 is a calcareous clayey loamy soil, characterized by alkaline pH as expected by its carbonate content of 55 g kg1. S2 is a loamy forest soil, characterized by acidic pH which is typical for these types of soil. The carbonate content in S2 was not measured since, at a soil pH lower than 7, it is considered negligible. The total organic carbon (TOC) content is considerably high (21.8 g kg1) in S2, whereas it is lower (7.7 g kg1) in S1. Although antibiotics are not usually detected in forest soils, S2 has been selected for its speciﬁc characteristics as low pH and high organic matter. P1 and P2 degradation was studied in both soils at wilting point and ﬁeld capacity. Wilting point is descriptive of soils at the highest redox potential due to maximal oxygenation conditions but is also due to severe dryness for soil biota (water content of 2.3% and 1.5% soil DW for S1 and S2, respectively). On the other hand, ﬁeld capacity is typical of higher amounts of water retainable by soils and thus available for biota (water content of 28.0% and 36.6% soil DW for S1 and S2, respectively). The two soil potentials have been chosen as ‘‘worst cases’’ for the degradation of this biotic type in that the severe dryness and low oxygen level at wilting point and ﬁeld capacity, respectively, make biota life difﬁcult. Hence these two extreme soil water potentials represent, amongst all possible others, the conditions of longest persistence for the two antibiotics under investigation in a soil environment. Antibiotic degradation at both water content conditions was also observed in soils after the reduction of microbial activity in order to discriminate the contribution of abiotic components from the degradation process. The persistence of P1, P2, and amoxicillin in S1 and S2, both untreated and autoclaved, at wilting point and ﬁeld capacity is reported in Fig. 3. Degradation kinetics data were processed, when 0 possible, using the ﬁrst order rate law (see kobs and t1/2, Table 3). For kinetics that did not follow the ﬁrst order rate law, t1/2 determination was done by visual interpolation of experimental data.
3.3.1. Antibiotics persistence at wilting point P1 degradation kinetics in (untreated) S1 and S2 soils at wilting point followed a ﬁrst order kinetics with t1/2 of 35.0 and 70.7 d, respectively (R2 P 0.8616, Table 3). In autoclaved soils, P1 persistence was higher than in untreated soils, owing to a severe reduction in microbial contribution to overall degradation, with t1/2 being slightly higher in S1 than in S2 (151.0 and 125.0 d, respectively). Conversely to the high degradability of P1 under alkaline aqueous conditions (t1/2 = 0.87 d at pH 9, Table 1), autoclaved calcareous soil S1 with pH 8.2 was able to degrade P1 less rapidly than autoclaved forest soil S2 with pH 5.0. It is most likely that this soil contains speciﬁc components which are able to catalyze P1 degradation. S2 is characterized by a consistent clay and total organic carbon (TOC) content whose amounts were roughly double those of S1: these components were probably more active in P1 degradation than the contribution of the alkaline catalysis. P2 degradation in untreated soils at wilting point did not follow a ﬁrst order kinetics and its t1/2 was visually deduced by the degradation curve reported in Fig. 3 on the left. The degradation was very quick (t1/2 = 1.7 and 2.5 d for S1 and S2, respectively), with a persistence that was slightly lower in S1 than in S2. As expected, P2 degradation in autoclaved soils was much slower than in the untreated ones and followed a ﬁrst order kinetics (R2 P 0.9236) with t1/2 of 32.2 d in S1 and 21.7 d in S2, conﬁrming the higher degradability of this antibiotic by soil microorganisms. In general, P2 resulted more degradable than P1 in both soils kept at wilting point.
I. Braschi et al. / Chemosphere 93 (2013) 152–159
Table 3 0 First order rate kinetic parameters (kobs = apparent kinetic constant; t1/2 = half life time) for P1 and P2 degradation in S1 and S2 soils at wilting point and ﬁeld capacity (both untreated and autoclaved) kept at RT. Parameters for amoxicillin are reported for comparison. Mean of three replicates; SD in parenthesis. Soil sample
P1 Untreated S1 Autoclaved S1 Untreated S2 Autoclaved S2
0.0201 0.0046 0.0100 0.0055
35.0 (6.2) 151.0 (9.3) 70.7 (14.0) 125.0 (4.8)
0.9779 0.9487 0.9772 0.8616
P2 Untreated S1 Autoclaved S1 Untreated S2 Autoclaved S2
– 0.0215 (0.0021) – 0.03195 (0.0022)
1.7* 32.2 (5.0) 2.5* 21.7 (1.5)
Amoxicillin Untreated S1 Autoclaved S1 Untreated S2 Autoclaved S2
n.a. n.a. n.a. n.a.
n.a. n.a. n.a. n.a.
(0.0035) (0.0003) (0.0019) (0.0021)
20.7650 (0.1626) – 10.9965 (0.2086) 0.0178 (0.0016)
0.03 (0.01) 0.70* 0.06 (0.00) 39.0 (3.6)
0.9941 (0.0060) – 0.9192 (0.0084) 0.9908 (0.0095)
– 0.9236 (0.0774) – 0.9547 (0.0118)
5.4610 0.8790 0.3927 0.2050
(0.4059) (0.0082) (0.0062) (0.014)
0.13 0.79 1.77 3.40
(0.09) (0.01) (0.03) (0.24)
0.9352 0.9498 0.9682 0.9880
(0.0602) (0.0021) (0.0397) (0.0150)
n.a. n.a. n.a. n.a.
1.6046 0.4101 1.2184 1.1748
(0.1485) (0.0953) (0.0045) (0.0262)
0.43 1.74 0.57 0.59
(0.04) (0.40) (0.00) (0.01)
0.9951 0.9924 0.9906 0.9923
(0.0055) (0.0089) (0.0018) (0.0080)
(0.0021) (0.0107) (0.0137) (0.0057)
Determination done by visual interpolation of experimental data as related data points do not follow a ﬁrst order kinetics. n.a.: Not available as no decreasing was found in 3 weeks observation.
Fig. 4. Degradation pathway of P1 and P2 in S1 and S2 soils.
As far as concerns amoxicillin degradation, no concentration decrease was detected in 3 weeks’ observation: therefore amoxicillin resulted as the most persistent among the selected b-lactams at wilting point. Amoxicillin unreactivity in soils at these conditions can be related to its physical state (i.e. powder) which makes its contact with solid soil components, which are only slightly hydrated at wilting point, very difﬁcult. On the other hand, P1 and P2 were added to soils in oily form and might pass bacteria membranes more easily even in the presence of such limited water content. This hypothesis is supported by the fact that P2, which is more hydrophobic than P1 (see CLog Kow, Fig. 1), was degraded faster than P1 in both untreated soils at wilting point.
3.3.2. Antibiotics persistence at ﬁeld capacity P1 persistence at ﬁeld capacity (Fig. 3, and Table 3) resulted very low in both untreated soils (t1/2 = 0.03–0.06 d) and autoclaved soil S1 (t1/2 = 0.70 d) whereas, in autoclaved S2, P1 was more persistent (t1/2 39.0 d), showing an important biotic contribution to the degradation process only for the acidic soil S2 at this water capacity. On the contrary, its low persistence in autoclaved S1 can be mainly ascribed to the catalytic effect of the alkaline pH of the soil solution. Compared to the wilting point, the soils at ﬁeld capacity resulted more active in degrading P1, indicating that the soil water potential seems to largely affect P1 degradation whereas the biotic contribution to the overall degradation only appears to be important in general in forest soil at wilting point.
P2 degradation at ﬁeld capacity was extremely fast in both calcareous soil S1 (untreated and autoclaved with t1/2 = 0.13–0.79 d) and forest S2 (untreated and autoclaved with t1/2 = 1.77–3.40 d). Comparing these results with those obtained at wilting point, it is clear that the contribution of the abiotic component to overall degradation is more relevant at higher soil water potential than at lower one. In fact, biotic and abiotic contributions to P2 overall degradation resulted similar at ﬁeld capacity. The persistence of amoxicillin in both untreated soils at ﬁeld capacity resulted very low as well, with t1/2 in the range of 0.43– 0.57 d and with a consistent contribution of the abiotic component to overall degradation (t1/2 = 1.74 and 0.59 d for autoclaved S1and S2, respectively). Unfortunately, no data for amoxicillin persistence/degradation in soils are available in the literature. Comparing our data for the reference b-lactam to those of P1 and P2, it is possible to assert that all the investigated antibiotics showed similar and very short persistence in soils at the highest soil water potential, whereas at wilting point they presented different stability (higher for P1 and lower for P2). 3.4. Degradation pathway in soils P1 and P2 degradation products monitored in soils can be visualized in Fig. 4. Two P1 degradation products in S1 at wilting point were detected and identiﬁed as P1-1 and P1-3 (already found in alkaline solution, Fig. 2). The formation of these products can be safely ascribed to the alkaline pH of soil S1, given that they are also
I. Braschi et al. / Chemosphere 93 (2013) 152–159
present in autoclaved soil. At ﬁeld capacity, the formation of P1-2 and P1-3 was readily observed at day 1 after the addition of the antibiotics, while at day 2 only P1-3 could be UV detected. Two additional by-products, P1-5 and P1-6 (structures not deﬁned) were found in S1 at wilting point, which was most likely due to microbial transformation. At ﬁeld capacity, no detectable P1 byproducts were found. Owing to the fast degradation of P1 in this soil at ﬁeld capacity (t1/2 0.03 d), predominant microbial degradation pathways can be hypothesized. In acidic soil S2, no UV–visible P1 degradation by-products were detected at both water potentials. Nevertheless, looking at the P1 degradation trend reported in Fig. 3, it is clear that biotic mechanisms were very active in reducing its concentration at wilting point whereas the abiotic component only gained a certain importance at ﬁeld capacity. Only two by-products (P1-7 and P1-8, structures not deﬁned) formed over time in autoclaved soil samples at wilting point. Their absence in untreated soil could be indicative of high degradability in the presence of biotic and/or abiotic alternative degradation pathways. Despite low P2 persistence in calcareous soil S1 (t1/2 0.13– 1.77 d), no UV by-products were detected, this is likely to be due to very fast biotic and/or abiotic transformations. In forest soil, P2-1 was monitored from the very ﬁrst contact hours along with two new by-products (P2-2 and P2-6). As the presence of all the three P2 by-products was also observed in the autoclaved soil, their formation can be ascribed to abiotic processes.
4. Conclusions P1 and P2 were found to be less stable than amoxicillin in water in the entire pH range with a few exceptions: P1 hydrolysis was catalyzed at acidic pH whereas P2 hydrolysis was at both acidic and alkaline pH values. P1 and P2 persistence in both soils was lower than what has been observed for amoxicillin at wilting point and in the same order of magnitude at ﬁeld capacity. P1 persistence in soils was negatively correlated with the soil water content: higher at wilting point (t1/2 151.0–125.0 d) and lower at ﬁeld capacity (t1/2 0.03–0.06 d). The low P1 stability in soils at ﬁeld capacity was mainly ascribed to the alkaline pH of the soil solution in calcareous soil and to microbial activity in forest soil. P2 persistence in soils was very low regardless of soil water potential (0.13–2.5 d) due to its high CLog Kow value (3.8789) so this molecule can easily pass bacterial membranes and, as a consequence, is more widely metabolized even with the low soil water content at wilting point. Three P1 hydrolysis products were identiﬁed in calcareous soil whereas a single P2 hydrolysis product was detected in forest soil. The formation of other P1 and P2 degradation products of microbial origin in both soils formed over time indicating an additional biotic degradation alternative to hydrolysis. In the light of these ﬁndings, it is possible to conclude that the newly synthesized monocyclic lactams P1 and P2 present a persistence in waters and soils which can in general be considered of lower ecosystem impact to that which has been observed for amoxicillin.
Acknowledgments Financial support was provided by: the University of Bologna (Strategic Research Project: A multidisciplinary approach to antibiotic resistance: new environmentally friendly antibiotics) and Fondazione Fibrosi Cistica onlus. Prof. Luca Calamai is acknowledged for his fruitful discussion. References Alef, K., 1995. Sterilization of soil and inhibition of microbial activity. In: Nannipieri, P., Alef, K. (Eds.), Methods in Applied Soil Microbiology and Biochemistry. Academic Press, San Diego, pp. 52–54. Bergheim, M., Helland, T., Kallenborn, R., Kummerer, K., 2010. Benzyl-penicillin (Penicillin G) transformation in aqueous solution at low temperature under controlled laboratory conditions. Chemosphere 81, 1477–1485. Broccolo, F., Cainelli, G., Caltabiano, G., Cocuzza, C.E.A., Fortuna, C.G., Galletti, P., Giacomini, D., Musumarra, G., Musumeci, R., Quintavalla, A., 2006. Design, synthesis and biological evaluation of 4-alkyliden-beta lactams: new products with a promising antibiotic activity against resistant bacteria. J. Med. Chem. 49, 2804–2811. Cainelli, G., Galletti, P., Garbisa, S., Giacomini, D., Sartor, L., Quintavalla, A., 2003. 4Alkylidene-azetidin-2-ones: novel inhibitors of leukocyte elastase and gelatinase. Bioorg. Med. Chem. 11, 5391–5399. Cainelli, G., Galletti, P., Garbisa, S., Giacomini, D., Sartor, L., Quintavalla, A., 2005. 4Alkyliden-beta-lactams conjugated to polyphenols: synthesis and inhibitory activity on matrix proteinases. Bioorg. Med. Chem. 13, 6120–6132. Black, C.A., 1965. Methods of Soil Analysis: Part I – Physical and Mineralogical Properties. American Society of Agronomy, Madison, Wisconsin, USA. Chada, R., Kashid, N., Jain, D.V.S., 2003. Kinetic studies of the degradation of an aminopenicillin antibiotic (amoxicillin trihydrate) in aqueous solution using heat conduction calorimetry. J. Pharm. Pharmacol. 55, 1495–1503. Conyers, M.K., Davey, B.G., 1988. Observations on some routine methods for soil pH determination. Soil Sci. 145, 29–39. Galletti, P., Quintavalla, A., Ventrici, C., Giannini, G., Cabri, W., Penco, S., Gallo, G., Vincenti, S., Giacomini, D., 2009. Azetidinones as zinc-binding groups to design selective HDAC8 inhibitors. Chem. Med. Chem. 4, 1991–2001. Galletti, P., Giacomini, D., 2011. Monocyclic b-lactams: new structures for new biological activities. Curr. Med. Chem. 18, 4265–4283. Galletti, P., Cocuzza, C.E.A., Pori, M., Quintavalla, A., Musumeci, R., Giacomini, D., 2011. Antibacterial agents and cystic ﬁbrosis: synthesis and antimicrobial evaluation of a series of N-thiomethylazetidinones. Chem. Med. Chem. 6, 1919– 1927. Dell’Aica, I., Sartor, L., Galletti, P., Giacomini, D., Quintavalla, A., Calabrese, F., Giacometti, C., Brunetta, E., Piazza, F., Agostini, C., Garbisa, S., 2006. Inhibition of leukocyte elastase, polymorphonuclear chemoinvasion, and inﬂammationtriggered pulmonary ﬁbrosis by a 4-alkyliden-beta-lactam with a galloyl moiety. J. Pharmacol. Exp. Ther. 316, 539–546. ESAC – European Surveillance of Antimicrobial Consumption, ‘‘Final Management Report 2009-2010’’. ECOC – European centre for Disease Prevention and Control ‘‘Annual epidemiological report on communicable diseases in Europe 2010’’ to be found under . Hou, J.P., Poole, J.W., 1971. b-Lactam antibiotics: their physicochemical properties and biological activities in relation to structure. J. Pharm. Sci. 60 (4), 503–532. Kümmerer, K., 2009a. Antibiotics in the aquatic environment – a review – Part I. Chemosphere 75, 417–434. Kümmerer, K., 2009b. Antibiotics in the aquatic environment – a review – Part II. Chemosphere 75, 435–441. Loeppert, R.H., Suarez, D.L., 1996. Carbonate and gypsum. In: D.L. Sparks (Ed.), Methods of soil Analysis. Part 3, Chemical Methods. SSSA Book Series no. 5 ASA and SSSA, Madison, WI, USA). Riesenfeld, C.S., Goodman, R.M., Handelsman, J., 2004. Uncultured soil bacteria are a reservoir of new antibiotic resistance genes. Environ. Microbiol. 6, 981–989. Sheldrick, B.H., Wang, C., 1993. Particle size distribution. In: Carter, M.R. (Ed.), Soil Sampling and Methods of Analysis. CSSS Lewis Publisher, Boca Raton, Florida, USA. Trevors, J., 1996. Sterilization and inhibition of microbial activity in soil. J. Microbiol. Methods 26, 53–59. Turos, E., Long, T.E., Konaklieva, M.I., Coates, V., Shim, J.Y., Dickey, S., Lim, D.V., Cannons, A., 2002. N-Thiolated b-lactams: novel antibacterial agents for methicillin-resistant staphylococcus aureus. Bioorg. Med. Chem. Lett. 12, 2229–2231. Walsh, C., 2000. Molecular mechanisms that confer antibacterial drug resistance. Nature 406, 775–781. Walsh, C., 2003. Antibiotics: Actions, Origins, Resistance. American Society for Microbiology (ASM) Press, Washington, DC. Wuts, P.G.M., Greene, T.W., 2007. Greene’s Protective Groups in Organic Synthesis, fourth ed. Wiley-Interscience Hoboken, New Jersey. pp 196–206.