| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Urban Water Management, Swiss Federal Institute for Aquatic Science and Technology (Eawag), Überlandstrasse 133, 8600 Dübendorf, Switzerland
b Dep. of Environmental Chemistry, Swiss Federal Institute for Aquatic Science and Technology (Eawag), Überlandstrasse 133, 8600 Dübendorf, Switzerland
* Corresponding author (christian.stamm{at}eawag.ch)
Received for publication September 7, 2006.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: A-SDM, N4–acetyl-sulfadimidine • BB, Brilliant Blue • LOD, limit of detection • LOQ, limit of quantification • SA, sulfonamide antibiotic(s) • SDM, sulfadimidine • SDZ, sulfadiazine • STZ, sulfathiazole
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
For the further fate of VA and their impact in the environment, sorption to soil is of crucial importance. Besides degradation it determines to what extent VAs remain available for transport (e.g., to water bodies) and biological uptake. Sorption properties vary strongly between the various groups of VAs (Tolls, 2001; Thiele-Bruhn, 2003). Tetracyclines and fluoroquinolones have been shown to sorb very strongly to soils rendering them rather immobile except for particle-bound transport. Sulfanomides are expected to be substantially more mobile due to their physicochemical properties (Tolls, 2001). This expectation has partially been confirmed by batch experiments (Langhammer, 1989; Boxall et al., 2002; Thiele-Bruhn et al., 2004) whereas other studies revealed a substantial sorption with longer contact times (Stoob et al., 2006; Kahle and Stamm, 2007). Because SAs also account for a large percentage of total VA use in European countries (20 to 80%) (LUA, 1999; Winckler and Grafe, 2001; Arnold et al., 2004), they are of special environmental interest and we have focused our experiments on this compound class.
Sulfanomides are excreted by animals to a large percentage either as parent compounds or as acetyl and hydroxy conjugates (Vree and Hekster, 1985; Böhm, 1996). In pig manure, concentrations of several SAs have been reported to range between 0.3 and 60 mg L–1 and reaching up to 20 mg L–1 for their corresponding acetyl-metabolites (Langhammer, 1989; Haller et al., 2002; Höper et al., 2003; Engels, 2004). Sulfadimidine (SDM, synonymous to sulfamethazine), and sulfadiazine (SDZ) were determined in 53 and 29% of manure samples (total n = 344 samples) taken in German pig farms (Engels, 2004).
The SA reaching the soil may persist for several weeks or even months in relatively high concentrations. In the top 5 cm of a grassland soil, 400 to 600 µg kg–1 SDM, SDZ, and sulfathiazole (STZ) were extracted 1 d after a regular application of pig manure (Stoob, 2005). Three months later, 15 to 20% of the applied amounts were still extractable by a harsh pressurized liquid extraction (Stoob et al., 2006). However, the pore water concentrations decreased more rapidly partially depending on pH. Sulfanomide may be present as cations, neutral species, or anions. First, due to the basic pH of manure the speciation in soil is shifted toward the anionic form rendering the SA much more mobile (Kahle and Stamm, 2007). Later, in contact with soil the neutral form, which sorbs much more strongly compared with the anion, is the dominant species.
Phenomenologically, Burkhardt et al. (2005) have confirmed this explanation in that SA concentrations in runoff from manured plots were significantly higher than from aqueous controls. It was shown that, on the one hand, manure decreased SA sorption due to the high pH of manure as well as due to the limited contact with sorbing soil surfaces. On the other hand, runoff volume on grassland increased due to particulate matter in manure sealing the soil surface.
It could be expected that the presence of manure exerted comparable effects on SA availability in the soil pore water. Since manure has been shown considerable influence on dynamics and amount of preferential flow in grassland soils (Stamm et al., 2002) one might further expect an impact on the vertical SA distribution in the soil pore water. Such vertical leaching could be relevant for ground water pollution and for an effect on microbial activity along preferential pathways. Data on this topic are scarce beside a few studies demonstrating preferential flow of SA into subsoil (Höper et al., 2003; Hamscher et al., 2005), to tile drains (Boxall et al., 2002, Weiß et al., 2006) and ground water (Hirsch et al., 1999; Hamscher et al., 2005).
Hence, the objective of this plot study was to investigate to what degree manure and its contact time with the vegetated soil surface affect the vertical distribution of SA in pore water of a grassland soil. It could be expected that basic manure would reduce sorption to soil compared with SA applied in water, causing increased SA concentrations in soil pore water and enhancing losses via preferential flow. We tested the hypothesis of a changed vertical SA distribution by comparing the concentrations of three SAs in pore water at different depths of manured and control plots with which SA were applied in aqueous solution. The conservative tracer bromide and the slightly sorbing tracer Brilliant Blue were included to compare physical and chemical processes controlling the vertical transport of the applied SA.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The soil (Table 1) is classified as a Eutric Cambisol (FAO) and had an initial water content close to field capacity before the start of the study. Physical properties of the soil are described by Wehrhan et al. (2007). Due to strong similarities in grassland management, soil properties, and climatic conditions preferential flow is expected as observed by Stamm et al. (2002).
|
|
Application to Plots and Soil Sampling
We applied 3 L m–2 of spiked liquid pig manure or aqueous solution to the corresponding plots using a watering can. One d or 3 d after application, rainfall was simulated by sprinkler irrigation. Hence, one plot was used for each of the combinations of two treatments (manure, control) and two contact times (1 d, 3 d). The irrigation amount was 30 mm and the intensity 20 mm h–1 corresponding to a heavy rainstorm in this area. Further details regarding the sprinkling procedure are described in Burkhardt et al. (2005).
The soil sampling was performed 1 d after the irrigation using a drilling system (Humax-Drilling Systems, Luzerne, Switzerland). At seven fixed positions soil cores (diameter = 6 cm) were taken down to a depth of at least 50 cm. Undisturbed soil cores (n = 10) of every horizon were taken for determining the bulk density. The soil samples were kept cool at 4°C and were frozen in the laboratory at –20°C.
Sample Preparation
Before the analysis of the spiked manure, the samples were diluted 1:5 with deionized water. Subsequently, the diluted manure was centrifuged at 3000 x g for 15 min at 10°C. The supernatant was filtered and diluted once more (overall 1:100), adjusted to pH 4.2, and filtered through one-way 0.45-µm membrane nylon filters again. Previous tests had revealed that no losses occurred when using these filters. The SAs were additionally extracted by pressurized liquid extraction (ASE 200 Accelerated Solvent Extractor; Dionex, Sunnyvale, CA). Details on the preparation of the manure samples and the extraction procedure are described in Burkhardt et al. (2005). The SAs were quantified by high pressure liquid chromatography coupled to electrospray ionization tandem mass spectrometry (HPLC-ESI-MS/MS) as described below.
To obtain pore water samples, the frozen soil cores were cut with a water-cooled diamond saw into slices of 2 cm thickness from the following depths: 0 to 2 cm, 5 to 7 cm, 10 to 12 cm, 20 to 22 cm, 30 to 32 cm, 40 to 42 cm, and 48 to 50 cm. The soil slices were thawed and 18 g of fresh soil were filled into centrifugation tubes. The tubes were centrifuged in an ultracentrifuge (Centrikon T 2000, Kontron Instruments, Switzerland) for 2 h with approximately 250000 x g (50000 rpm). The supernatant of 1 to 2 mL was taken with one-way pipettes and filtered through one-way 0.45-µm membrane nylon filters (Millipore, Bedford, UK). An aliquot of 200 µL of filtrate was transferred to a 2 mL amber HPLC vial. Isotope-labeled internal standards were added for each SA (SDZ, STZ: Toronto Research Chemicals, North York, ON, Canada; SDM: Cambridge Isotope Laboratories, Andover, MA) to account for effects of matrix suppression (Table 3). The prepared vials were stored at 4°C until analysis.
|
The water content of every soil sample was calculated by weight difference of the fresh soil and oven-dried soil (105°C, 24 h). To determine the initial water content, the same procedure was applied to soil samples taken directly adjacent to the irrigated plots. The bulk density of each horizon was calculated by weight difference of fresh and oven-dried soil (105°C, 24 h) in the undisturbed samples of known volume.
Analytical Procedures
Sulfonamides and their acetyl-metabolites were quantified in soil pore water using HPLC-ESI-MS/MS (HPLC: Rheos 2000, Flux Instruments, Switzerland; ESI-MS/MS: Finnigan TSQ Quantum with ESI, Thermo Electron Corporation, USA). The HPLC system included a reversed-phase C18-column (Luna C18, 150 x 2 mm, particle size 5 µm; Phenomenex, USA) coupled with a pre-column (ODE, Octadecyl, 4 x 2 mm, particle size 5 µm; Phenomenex, USA). Eluent A was methanol and eluent B water, which both were buffered to pH 4 by adding 40 µL of a 100% acetic acid to 1000 mL of eluent. The flow rate was set to 250 µL min–1, the injection volume was 25 µL. Details of the chromatographic separation with a gradient and validation procedure are described in Stoob et al. (2006) and Zimmermann (2006).
The SA, their metabolites, and the isotope-labeled internal standards were detected in the single reaction monitoring (SRM) mode using positive electrospray ionization (ESI+). While two product ions were acquired for quality control, the first product ion was used as the quantifier (Table 3). The LOD was defined by the first product ion quantifiable with a signal to noise ratio (S/N) of 10/1 and the second product ion detectable by S/N of 3:1. Limit of quantification (LOQ) was exceeded if both product ions were quantifiable with S/N of 10/1 and deviations of the qualifier from the quantifier were in the range of ± 10%. The LOD and LOQ in pore water was low and defined for each SA at 1 and 5 µg L–1, respectively, for the parent compounds. The corresponding values for the acetyl-metabolites were slightly higher with a LOD of 5 µg L–1 and a LOQ of 15 µg L–1. In the low concentration range we tested for individual samples whether they exceeded LOD or LOQ and mentioned this where appropriate.
For Br analysis ion chromatography (761 Compact IC; Metrohm, Herisau, Switzerland) coupled to a Metrosep A Supp 4 column and a pre-column (Metrohm) was used. For further details see Burkhardt et al. (2005). The BB measurement was performed with a photometer (Uvikon 940 Spectrometer, Kontron, Switzerland) at 630 nm wavelength.
Recovery Calculations and Sorption Coefficient
The relative concentration C% of a given solute was determined by normalization of the measured soil pore water concentration C to C0meas multiplied by 100 where C0meas corresponds to the measured input concentration taken 24 h after spiking the applied manure and water.
The total mass balance was estimated for each of the four plots to determine the recovery of the applied amount. The relative mass recovery, R% (%), was calculated for n = 7 samples per column and n = 7 columns [1].
 | [1] |
 | [2] |
i the soil bulk density, F the plot area, and dz the depth of the soil layer. In the second step, the mass of the solutes in the soil between the analyzed layers was calculated according to the same Eq. [2], but inserting the mean values for the measured pore water concentration, gravimetric water content, and the bulk density of the adjacent layers above and below.
The ratio of relative SA concentrations C%SA compared with those of Br C%Br allows an estimation of apparent distribution coefficients Kd,app in the topsoil (0–2 cm). Assuming linear sorption and negligible mass flow downward, Kd,app may be calculated as follows:
 | [3] |
From the difference of the nominal and the measured SA concentrations in the supernatant of manure it is possible to calculate the Kd,app as well. If normalized to organic carbon content of 50% in the solid phase of manure the KOC can be calculated.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
These results indicate that SA may sorb to manure constituents despite its high pH of 8.1. In our case, the Kd,app ranged from 30 to 40 L Kg–1 and KOC from 60 and 80 L Kg–1. Similar findings have been obtained in batch experiments for STZ in manure at the same pH (Kahle and Stamm, 2007).
In contrast to the SA, the measured concentrations of the two tracers (Br, BB) in the liquid phase of manure (C0measBr, C0measBB) were almost identical to the nominal values. In the following, we used the concentrations of all solutes measured in the supernatant as the initial concentration for the solute comparison.
Bromide Depth Distribution
For both treatments (manure, controls), Br concentrations decreased strongly with depth (Fig. 1). The median Br concentrations in the manured plots declined from >1300 mg L–1 at the soil surface to 100 mg L–1 at 10-cm soil depth, whereas the corresponding median concentrations in the control plot were 1000 and 370 mg L–1 (Fig. 1). This implies a tendency for a steeper depth gradient on manured plots compared with the control. However, the variability was so large that the observed differences were not statistically significant. In both treatments, Br was found down to a depth of at least 50 cm (maximum depth systematically sampled and analyzed). Given the irrigation amount of 30 mm this vertical Br transport clearly demonstrates preferential flow, a conclusion corroborated by the BB and SA results presented below. At the three largest depths, the Br concentrations for both treatments were rather similar (Fig. 1).
|
The median mass recovery of Br in the soil pore water ranged from 58.1 to 76.4% on each plot (Table 4). The large spatial variability mentioned above is reflected in standard deviations of 16.4 to 29.2% based on seven soil columns per plot. The recoveries were independent of application matrix (manure, aqueous solution) and contact time (1 d, 3 d). The recoveries demonstrate a reasonable representativeness of the soil sampling procedure and the efficiency of our soil pore water extraction method.
|
|
|
At greater depths, there was a tendency for more samples with higher concentrations in the controls. This was very pronounced for BB (Fig. 3), but could be found also for SDZ and STZ. For these two SA, only the controls had median concentrations > LOD below 0- to 2-cm depth. In contrast, no difference was observed between the maximum concentrations found in the different treatments. A possible explanation for this observation would be the sealing of smaller pores at the soil surface due to manure particles resulting in lower infiltration rates. This might be due to the sealing of parts of the preferential flow paths caused by manure and would indicate a higher flow concentration (Flühler et al., 1996) toward fewer but larger macropores. This mechanism would enhance the flow rate through those pores as has been observed in other experiments (Stamm et al., 2002; Weiler and Naef, 2003). The effect of particles covering soil pores favors ponding and surface runoff (Burkhardt et al., 2005).
The relative SA and BB concentrations normalized to those of Br (Fig. 3) allow a direct comparison of pore water data to that of runoff data presented in a previous paper (Burkhardt et al., 2005). This comparison clearly reveals that the medians of the SA concentrations in pore water were much lower relative to Br than in surface runoff. Whereas relative SDM concentrations ranged between 30 and 60% of the Br concentrations in runoff (Fig. 5 in Burkhardt et al. [2005]), these relative values were always below 10% in soil pore water even in the top layer. These findings indicate a substantial influence of sorption in soil on SA availability and they suggest that this effect is more pronounced for vertical transport compared with runoff. Grass cover interception might have a major influence on surface runoff of SA under field conditions right after manure application. When rainfall starts the solutes are washed off and routed directly into runoff before they have time to react with the soil. On the other hand, that portion of SA that gets into contact with the soil surface undergoes sorption which becomes increasingly important with time (Stoob, 2005; Hamscher et al., 2005; Kreuzig and Höltge, 2005; Kahle and Stamm, 2007). Hence, it is plausible that SA concentrations are higher in runoff (Burkhardt et al., 2005) due to a shorter contact time compared with pore water of soil cores that were taken 1 d later.
For the manured plots, the distribution coefficient Kd,app yielded values for SDM in the order of 1.8 kg L–1 after 1 d contact time and of 3 kg L–1 after 3 d contact time (KOC of about 80 kg L–1). For the two spiked SAs (SDZ, STZ) the values are about four times higher. These values are in the same order of magnitude found by Stoob (2005) for SA in the same soil for the 0- to 5-cm layer 1 d after manure application. In that study the comparison of pore water concentrations with SA concentrations after ASE extraction yielded apparent Kd,app values for SDM, STZ, and SDZ between 10 and 15 kg L–1. The Kd,app values on the control plots were higher reaching values of about 30 to 35 kg L–1 (KOC = 830–970 kg L–1).
The measured SDM concentrations in pore water were always similar to or higher than the concentrations of the spiked SAs (SDZ, STZ) (Fig. 2) although the measured input concentration of SDM C0measSDM was only one-third of that of the spiked SAs. This indicates that the affinity of SDM to soil surfaces was less pronounced compared with the other two SAs. It could be argued that this was an effect of SDM being administered to animals whereas STZ and SDZ were spiked to manure shortly before the experiment. Hence, it is possible that sorption of SDM has reached equilibrium in the manure whereas the sorption of the spiked SA will increase with time. However, batch experiments comparing SDM with the other SA confirmed a weaker SDM sorption to organic matter with the same contact time for all three SA (Zimmermann, 2006), a finding supported by results at the field scale in subsoil and ground water (Höper et al., 2003, Weiß et al., 2006).
The weaker sorption of SDM was also reflected in the recoveries of the different SAs. Whereas about 10 to 15% of the applied SDM was found in pore water, the recoveries for SDZ and STZ were about three times lower (Table 4). There was a clear trend to lower recoveries on the controls compared with the manured plots for the spiked SA.
Despite the fact that the plot experiments with manure described here and in the previous paper represent worst case situations, only <0.01 to <2% of SA were lost with surface runoff from manured plots and generally <7% of the SAs were available in soil pore water. The exception was SDM with up to 20% in pore water. For the entire soil profile, 87 to 99.9% of the applied SA were retained in the topsoil. This high retention holds for Br as well, demonstrating that the limited physical exchange between soil solution and the irrigation water in the uppermost soil layer was the main factor limiting the SA losses via runoff.
Formation of Sulfonamides Metabolites
So far, we have neglected—e.g., in the mass balance calculation—the fact that SA may be transformed to acetyl-conjugates that can be present in manure in substantial amounts (Langhammer, 1989; Haller et al., 2002; Weiß et al., 2006). In our case, however, we could not detect any acetyl-metabolites in the manure samples above the LOD, which was about 100 µg L–1 in the manure. Therefore, this metabolite did not exceed 1% of the parent compound. In contrast to that, we found acetyl-SDM (CA-SDM) in all pore water samples of the uppermost soil layer (0- to 2-cm depth) after manure application and irrigation. The CA-SDM concentrations varied between 3 and 18 µg L–1 corresponding to a proportion of 10 to 30% of the parent CSDM. Despite the fact that some of the concentrations were between the LOD and LOQ, the detections in all samples correspond to 10 to 30 fold higher concentrations than expected based on the metabolite concentration and the corresponding LOD, respectively, in the manure. An influence of contact time was not measurable. In other studies, the CA-SDM proportion was 10 to 80% in manure and drainage water (Langhammer, 1989; Engels, 2004; Stettler, 2004; Weiß et al., 2006). Wehrhan (2006) has investigated the occurrence of transformation products at different time steps after spiking fresh soil. One day after spiking two transformation products of SDZ, acetyl- and hydroxy-SDZ were identified in proportions of >20% for each and the concentration decreased with increasing time. Of the spiked SAs only acetyl-STZ (CA-STZ) was detectable at the soil surface (CA-STZ < 5 µg L–1), but at levels too low for reliable quantification (Wehrhan, 2006).
In contrast to the clear detection of acetyl-SDM in the top-most soil layer in our experiment this metabolite was not observed in any of the samples of the lower layers. Even in samples of comparable concentration of the parent compound as in the topsoil, no signal of acetyl-SDM was detected. It seems that the formation of metabolites was limited to contact of manure with the very topsoil.
Despite the fact that the acetyl-SDM concentrations have to be treated with some caution because they were partially between LOD and LOQ it is interesting to notice a statistically significant correlation between the CA-SDM and CSDM concentrations. The more parent compound the higher the metabolite concentration. However, ratio between metabolite and parent compound decreased with increasing CSDM. Similar findings were reported by Wehrhan (2006).
Besides the acetylated transformation products, other metabolites are known to be formed from SA (Kreuzig and Höltge, 2005; Wolters and Steffens, 2005; Wehrhan, 2006). These were not covered with our analytical method. Laboratory experiments with soil from our test site using radiolabeled SDZ (14C-SDZ) under sterile and nonsterile conditions clearly revealed three peaks of transformation products including acetyl- and hydroxy-SDZ and one unknown product (Wehrhan, 2006). Based on the peak intensities it was estimated that up to 50% of SDZ was transformed to new products within 1 d. It seems that the transformation was at least partially abiotic. In another study, Kreuzig and Höltge (2005) observed four unknown transformation products. These products were determined in the fresh manure-soil mixtures 3 d after spiking, but neither in pure manure nor in sterile soil samples. The results clearly indicate a fast, not identified transformation process at the soil surface. Other transformation processes like photodegradation have been reported in the literature (Wolters and Steffens, 2005) as well. Under our experimental condition with plots covered by plastic sheets and partially by wood panels just after the application and due to the strong light absorption in manure we consider this process to be negligible. This suggests that the actual distribution coefficients derived from the present study may overestimate the affinity of SA to soils since the formation of transformation products has not been taken in account.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Beside the effect on microbial communities in soil, SA may leach in agricultural soils to drainage systems and occur in surface waters. This has been actually shown by Stoob (2005) on the same grassland site. However, the contamination of such a loamy soil seem to be limited by fast sorption and transformation and therefore of temporary impact to surface water. Nevertheless, the farming system (pasture or tillage) is of main influence on sorption and potential losses to greater depths. It seems plausible that manure application on grassland may favor SA mobility compared with an application on plowed soil or injected into the top layer. Moreover, sorption and mobility of transformation products has to be taken into account for the risk assessment as long as the product is effective or may be retransformed to the parent compound.
The effect of different treatments on the vertical pore water SA content was limited. This might be due to the sealing of parts of the preferential flow paths by manure particles. Our results are in line with the observation that manure increased runoff volume compared with controls. Overall, the data from our study indicate that the different treatments (contact times, manure vs. aqueous application) had a strong influence on the load of SA lost from the plots by fast transport processes. In contrast, the depth distribution of resident concentration in pore water, which is relevant for possible biological effects, was much less affected by the treatments.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Larsbo, K. Fenner, K. Stoob, M. Burkhardt, K. Abbaspour, and C. Stamm Simulating Sulfadimidine Transport in Surface Runoff and Soil at the Microplot and Field Scale J. Environ. Qual., May 1, 2008; 37(3): 788 - 797. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Soil Science Society of America Journal | Journal of Plant Registrations | The Plant Genome | |||
