Rainfall increases the abundance of antibiotic resistance genes within a riverine microbial community

Environmental Pollution xxx (2017) 1e6

Contents lists available at ScienceDirect

Environmental Pollution journal homepage: www.elsevier.com/locate/envpol

Rainfall increases the abundance of antibiotic resistance genes within a riverine microbial community* Andrea Di Cesare a, 1, Ester M. Eckert a, 1, Michela Rogora b, Gianluca Corno a, * a b

Microbial Ecology Group (MEG), National Research Council - Institute of Ecosystem Study (CNR-ISE), Largo Tonolli, 50, 28922, Verbania, Italy National Research Council - Institute of Ecosystem Study (CNR-ISE), Largo Tonolli, 50, 28922, Verbania, Italy

a r t i c l e i n f o

a b s t r a c t

Article history: Received 19 December 2016 Received in revised form 9 March 2017 Accepted 17 April 2017 Available online xxx

Infections with antibiotic resistant bacteria are among the major threats for human health. Studies elucidating the role of the environment in their spread are still in their infancy, it, however, seems that different environments might function as a long-term reservoir of antibiotic resistance genes (ARGs) that reside within their microbial communities. An increasing number of studies target the presence and the persistence of ARGs in waters and soils that are exposed to human activities; they, however, rarely consider the spatial/temporal variability that predominate in these environments. Here we evaluated the effect of a moderate rain event (4 mm rain h
1) on the abundance and distribution of ARGs (tetA, ermB, blaCTXM, sulII, and qnrS), by comparing measurements of gene abundances during the rainfall to the yearly average, in the waters of a large subalpine river. ARG abundances, which all increased during the rain event, were then correlated to several microbiological, physical and chemical variables, in order to establish their potential origin. Increments in ARG abundances during rainfall (total ARGs: 24 fold) was concomitant to an increase in total phosphorous, N-NH4, and microbial aggregates. Our results show a strong influence of a moderate rainfall on the abundances of ARGs, and suggest the catchment as their source. The impact of moderate rainfalls in areas exposed to anthropic activities should then be considered in modelling and management of ARG dynamics. © 2017 Elsevier Ltd. All rights reserved.

Keywords: Antibiotic resistance River pollution River microbial community Rain Anthropic impact

1. Introduction The spread and persistence of antibiotic resistant bacteria (ARB) and antibiotic resistance genes (ARGs) in the environment is a major threat for the management of water and land resources with a strong influence on human health (Berendonk et al., 2015). Antibiotic resistances are common in bacterial communities in nature (Cytryn, 2013; Czekalski et al., 2015; Di Cesare et al., 2015), and the relative proportion of resistant bacteria seems to be correlated to human activities in lakes, rivers, soils, and oceans (Amos et al., 2014; Di Cesare et al., 2015; Hatosy and Martiny, 2015; Popowska et al., 2012; Pruden et al., 2012). During the last decade a number of studies (summarized by Berendonk et al., 2015) highlighted the presence of various ARGs (e.g. genes sulI, sulII, and sulIII

* This paper has been recommended for acceptance by Dr. Harmon Sarah Michele. * Corresponding author. E-mail address: [email protected] (G. Corno). 1 These authors contributed equally to the article.

against sulphonamides; tetA, tetW, and others against tetracyclines, blaTEM, blaCTXM, and others against beta-lactams; qnrS against quinolones; ermB against macrolides) within the microbiome of water bodies. Occasionally, even the presences of ARGs against last resource antibiotics such as vancomycin (Schwartz, 2012) and colistin have been found (Zurfuh et al., 2016). In fact, specific resistances against synthetic and semisynthetic antibiotics persist in the environment, where water bodies act as long-term reservoir for ARGs (Di Cesare et al., 2015; Hatosy and Martiny, 2015), which might hamper the success of measures taken to reduce the prevalence of resistant bacteria (Taylor et al., 2011). The most powerful techniques applied for the assessment of the presence and abundance of ARGs within bacterial communities in the environment (e.g. quantitative Real Time-PCR, high throughput meta-sequencing) provide detailed community datasets (Berendonk et al., 2015), that allow statistical correlations between different genomic fragments (e.g. co-occurrence of ARGs and heavy metal resistance genes, Di Cesare et al., 2016a) and between ARGs and specific bacterial strains or other ecological variables (Ju et al., 2016).

http://dx.doi.org/10.1016/j.envpol.2017.04.036 0269-7491/© 2017 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Di Cesare, A., et al., Rainfall increases the abundance of antibiotic resistance genes within a riverine microbial community, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.04.036

2

A. Di Cesare et al. / Environmental Pollution xxx (2017) 1e6

An increasing number of such studies assessed temporal and spatial variations in the distribution of ARGs, demonstrating high seasonal (Di Cesare et al., 2015) and spatial (water surface, water column, and sediments; Czekalski et al., 2012) variability, and the importance of ecological niches (interactions between bacteria, potential symbiosis with larger organisms) (Czekalski et al., 2014; Eckert et al., 2016). Other studies measured the impact of socalled hot-spots of ARGs like wastewater treatment plants (WWTPs) (e.g. Di Cesare et al., 2016b), urban areas and agricultural soils (Popowska et al., 2012; Pruden et al., 2006). Unfortunately, the largest majority of these studies are based on punctual sampling, or on series that do not consider small or moderate meteorological disturbances. In some cases, weather events that are outside the norm are intentionally excluded from the sampling campaign in order to ensure higher reproducibility of the data over time (Di Cesare et al., 2015). Disturbances, however, might have a sudden and significant impact on the ARG abundances, which might be missed when avoiding such events. One potential disturbance that is very common in temperate areas are moderate rain events (with rain intensity peaking between 2.5 and 7.6 mm rain d
1; American Meteorological Society, 2000), which could influence the overall ARG load in a water body by modifying the composition of the allochthonous bacterial load, e.g. through increased input from the catchment area (Amos et al., 2015). In addition, the nutrient composition of the water body might change during such events since runoff originating after rainfall is one of the main pathways for nitrogen and phosphorus compounds from the catchment to rivers and lakes (Hathaway et al., 2012; Janke et al., 2014), particularly after prolonged dry periods (Chen et al., 2016; Outram et al., 2014). This study provides the first investigation of the impact of a common disturbance, a moderate rain event, to the resistome of a large subalpine river. We assessed the abundance of five different ARGs (tetA, ermB, blaCTXM, sulII, and qnrS) in the River Toce (Italy), that confer resistance to different classes of antibiotics and were previously confirmed to be often present and abundant in Lake Maggiore and/or the surrounding WWTPs (Di Cesare et al., 2015; 2016b,c). We then related the gene abundances to classical microbiological variables (bacterial number and phenotypical distribution) and to selected chemical parameters (Total Phosphorus (TP), ammonium (N-NH4) and TOC). Our results allow (1) the evaluation of the impact of the event itself in comparison to a 1-year long measurement of the ARGs load of the River, and (2) speculations of the potential origin of the ARGs during the rain event. 2. Material and methods 2.1. Site description and sampling The River Toce is located in the Central Alps, in North Western Italy, Piedmont region. A limited part of its watershed is located in Switzerland (10%). It is the second largest tributary of Lake Maggiore after the River Ticino. The Toce watershed is a typical glacial basin with steep hill slopes bounding a narrow valley. The total drainage area is 1800 km2. Its altitude varies between 193 m a. s. l. at the outlet to the 4600 m a. s. l. of the highest peak of Mount Rosa (Ravazzani et al., 2016), with 40% of the basin area higher than 2000 m. Irrigation is not present in the watershed and drinking water uses are negligible (Ravazzani et al., 2016). The total population in the watershed is around 122.000 inhabitants (61 inhabitants km
2), about 75% of which is directly connected by sewage to the treatment plants located in the Toce catchment area. The percentage of catchment area devoted to agricultural practices is rather limited (6.7%); most of the catchment is indeed covered by forest, shrubland and grassland.

The total loads of total phosphorus (TP) and nitrogen (TN) delivered to River Toce have been estimated as about 100 t P y
1 and 1500 t N y
1, respectively. In the case of TP, the contribution of point and diffuse sources is similar (45 and 55%, respectively), while for TN the contribution of non-point sources such as agriculture, livestock and atmospheric deposition is much higher (almost 80% of the total). Three main WWTPs release their effluents into the River Toce, namely, from the estuarine (those listed by proximity to our sampling site): Gravellona Toce WWTP (18000 p.e., 45 560 1700 N, 8 260 0200 E), Domodossola Sud (16,000 p.e., 46 050 3200 N, 8 170 4600 E), Domodossola Nord (16,000 p.e. 46 060 1200 N, 8 180 0900 E). In total, they release about 9  106 m3 of treated wastewater per year in River Toce. The overall quality status of River Toce water is good, as testified by the long-term monitoring of this river chemistry (National Research Council of Italy - Institute of Ecosystem Study, 2015). In order to estimate the annual abundance of ARGs, the surface water samples (around 1L) were recovered every two months from July 2015 to May 2016 just before the Toce estuarine in Lake Maggiore, at coordinates 45 560 0900 N, 8 260 4000 E. To evaluate the effect of the rain event on the dynamics of ARGs, water samples were taken on February 6th before the rain, on 7th every 3 h from 10:00 AM to 10:00 PM during the rain event. Meteorological data (hourly precipitation) at the station of Candoglia (45 580 3200 N, 8 250 2200 E) were provided by the Regional Agency for Environmental Protection (ARPA Piemonte). 2.2. Chemical variables On each sample, we determined nutrient concentrations (phosphorus and nitrogen compounds) by UV-VIS spectrophotometry and total organic carbon (TOC) by high temperature catalytic combustion. The analysis were completed as soon as possible after the sample collection, possibly within 1e2 days. The same variables are also routinely analyzed on River Toce samples on a monthly basis (National Research Council of Italy - Institute of Ecosystem Study, 2015). All the analyses were performed according to Standard Methods (APHA et al., 2012) and the QA/QC procedures in use in the laboratory. Details on the analytical procedures and the quality control can be found at http://www. idrolab.ise.cnr.it/en/. For data analysis, we tested TP, N-NH4, and TOC since they were assumed to be representative of P, N, and organic C forms stored in the soils, which can be delivered from the catchment to the river water during rain events (Outram et al., 2014; Lee et al., 2016). The data analyses aimed to assess (1) the influence of the rain event on the patterns of these chemical variables in river water and, (2) the potential relationships between chemical parameters and ARGs, suggesting a possible common source. 2.3. Bacterial abundance and size distribution Aliquots of each sample were prepared for the bacterial abundance and bacterial size distribution as described in Di Cesare et al. (2015). The bacterial numbers and the phenotypic distribution (single cells and small microcolonies; large aggregates composed by at least 10 clustered cells) were assessed for each sample by flow cytometry (Accuri C6, BD Biosciences, USA). The diverse phenotypes were assessed in order to understand the possible origin of the bacterial assemblages, or changes in nutrients composition in the water. In both cases, shifts towards larger cell aggregates or single free living cells can be detected as response of the bacterial community to modified composition or ecological factors (Corno and Jürgens, 2008). In detail, 0.5 mL for each liquid sample, fixed in formalin (final concentration 1%) were stained with SYBR Green I

Please cite this article in press as: Di Cesare, A., et al., Rainfall increases the abundance of antibiotic resistance genes within a riverine microbial community, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.04.036

A. Di Cesare et al. / Environmental Pollution xxx (2017) 1e6

(Life Technologies, USA) solution (1%) for 15 min in the dark. Counts (on triplicate samples) were set to a minimum of 2  106 events within the gate designed for single bacterial cells, and 5  102 events in the gate designed for bacterial aggregates (Corno et al., 2013). The correct assessment of flow-cytometric counts was confirmed by a preliminary check of single cells and aggregate numbers for 6 DAPI stained samples by epifluorescence microscopy
 (Axioplan; Zeiss, Germany) (Horn ak and Corno, 2012). 2.4. DNA extraction and 16SrDNA and ARGs quantification The water samples were pre-filtered on 10 mm net, collected on 0.2 mm polycarbonate filters (300e500 mL, Millipore), and stored at
20  C until molecular analysis. The filters were cut in two sections and were processed for the DNA extraction as previously described (Di Cesare et al., 2015). Every DNA sample was 2 fold diluted and tested for the abundance of 16SrDNA gene and for tetA, ermB, blaCTXM, sulII, and qnrS by qPCR (gene selection criteria summarized in Di Cesare et al., 2015, 2016c). All qPCR assays were performed using the RT-thermocycler CFX Connect (Bio-Rad Laboratories Inc., USA) following amplification programs and primers selection as described in Di Cesare et al. (2016b). The specificity of qPCR assays was evaluated by analysis of melting profile and electrophoresis, while the standard curves were performed following Di Cesare et al. (2013). The limits of detection (LOD) for all the standard curves were determined according to Bustin et al. (2009). In detail, 16SrDNA, tetA, ermB, sulII, blaCTXM, and qnrS showed a LOD of 58, 11, 54, 41, 73, and 62 copies mL
1, respectively. The potential inhibition of the matrix has been evaluated by the dilution method (Di Cesare et al., 2013), and no inhibition was observed. The average ± standard deviation of the efficiency and of R2 for all the runs were 100.55 ± 8.66 and 0.98 ± 0.02 respectively. The interpretation of values from the two replicated measurements were evaluated as described in Di Cesare et al. (2015). The abundances of each ARG are expressed as absolute values (copies of ARG mL
1) for the monthly monitoring and for the rain event, while they were expressed as relative values (copies of ARG per copy of 16SrDNA) during the rain event only.

3

cell and aggregate numbers, total ARG copies per ml and the relative abundances of each gene. Normality of the data was tested and the factors were log transformed if necessary. In a second step, we tested whether total absolute gene abundances could be explained by TP and N-NH4 (model 2) or by aggregate numbers (model 3). One-way repeated measure ANOVA with Tukey post-hoc test, conducted in SigmaPlot 12.5 (Systal Software, Inc.), were applied to determine whether the measured variables, namely TP, NH4, TOC, single cell and aggregate number and total ARG copy numbers per ml, were higher during the rain event compared to their yearly average (monthly sampling). 3. Results 3.1. Changes in prokaryotic abundances and chemical variables After 125 days without precipitations that excided 10 mm day
1 (the previous significant rain event dated October 4th, 2015, during the 125 days of draught only 22 mm of rain fell in the area, data

2.5. Data analysis Person correlations and linear models were conducted with R 3.1.2 (R Core Team, 2013) using RStudio Team (2015). First Pearson correlations were applied to evaluate whether the absolute gene abundances (gene copies per ml filtered water), including the sum of all the gene copies per ml, were correlated to each other, to their relative abundances (gene copies per 16S rDNA gene copy). Due to the strong correlation of all the absolute values of the genes per ml, further statistical analysis were performed using the sum of all the genes quantified (Table 1). We then applied linear models (LM) to assess the influence of rain (three-hour sum preceding sampling) on various parameters (model 1) including: TP, N-NH4, TOC, single Table 1 R2 of Pearson correlations of the absolute copy number of each ARG and the sum of all genes (tot ARG) with each other and their corresponding relative abundances (per 16S rDNA gene copy). Bold font highlights statistically significant R2 values. Gene ml
1 tetA ermB qnrS sulII tot ARG

tetA 0.926 0.950 0.907 0.948

ermB

0.993 0.978 0.997

qnrS

0.981 0.998

Gene 16S
1 0.541 0.666 0.631 0.740

Fig. 1. Dynamic of ARGs, chemical, and microbiological variables during the rain event. Amount of rainfall, absolute (A) and relative abundance (B) of the antibiotic resistance genes sulII, tetA, ermB and qnrS, concentrations of N-NH4, total phosphorous (TP) and total organic carbon (TOC) (C), abundances of single cells and aggregates (D) from 9AM on Feb. 6th 2016 to 11AM Feb. 28th 2016. The grey shaded area, which refers to the amount of rainfall, is repeated in all plots as reference. The rain event is marked with a grey line below the x-axis. Note the axis break in the y-axis.

Please cite this article in press as: Di Cesare, A., et al., Rainfall increases the abundance of antibiotic resistance genes within a riverine microbial community, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.04.036

4

A. Di Cesare et al. / Environmental Pollution xxx (2017) 1e6

provided by ARPA Piemonte), rain fall started around 10 PM on February 6th, 2016 (Fig. 1). Precipitation peaked between 1 and 3 PM on February 7th with values over 3 mm h
1 (Fig. 1). The rainfall stopped entirely after 10 PM that day. Before the rain, prokaryotic numbers were around 7.7  105 single cells and 1.8  103 aggregates ml
1. Both single cells and aggregates increased to maxima of 9.9  105 at 1 PM for single cells and 3.8  103 aggregates at 10AM on February 7th. Single cell and aggregate numbers then declined to reach numbers similar to before the rain with decreasing rain intensity around 10 PM. On the other hand, TP, TOC and N-NH4 increased during the rain event peaking at 51 mg P L
1, 0.99 mg C L
1 and 0.13 mg N L
1, respectively (with respect to 18 mg P L
1, 0.51 mg C L
1 and 0.04 mg N L
1 before the rain event) (Fig. 1). N-NH4 concentration and bacterial aggregates were significantly higher during the rain event compared to their yearly average (P-value < 0.05), whereas TP and TOC and single cell numbers did not significantly differ from their yearly average (Pvalue > 0.05, Table 2). However, the mean TP concentration during the rain event (42 mg P L
1) was considerably higher than the yearly average (26 mg P L
1; Table 2).

3.2. ARGs abundance during the year and during the rain event The abundances of the ARGs during the year were variable; tetA was the only always quantifiable ARG, with the highest value measured in January 2016 (4.59  102 copies mL
1) and the lowest in July 2015 (14.2 copies mL
1). The other ARGs (sulII, qnrS, ermB) showed the highest abundances in January 2016 and the lowest in the first three sampling dates (July, September, and November 2015) (Supplementary Table 1), while blaCTXM was never quantifiable but detectable in July and November 2015 and in January and March 2016. ermB was the gene with the highest variability in absolute abundances during the year (Supplementary Table 1). Looking at the abundances of ARGs during the rain event, tetA, sulII and ermB were always quantifiable, qnrS was always quantifiable except for the first time point (Fig. 1) while blaCTXM was only positive but not quantifiable, for this reason it was excluded from the statistical analysis. The ARGs absolute abundances were strongly correlated with each other (R2  0.90) during the rain event; they were however not significantly correlated to their respective relative abundances (Table 1). Before and after the rain event both the absolute and relative ARG abundances were lower than those measured during it (Fig. 1). The absolute abundances of ARG were significantly higher during the rain event compared to their yearly average (Pvalue < 0.05, Table 2). Table 2 Mean value (±standard deviation) of chemical and bacterial parameters and the sum of the absolute values of ARGs (tot ARG) during the rain event and of the monthly sampling during one year. F- and P-value of one-way repeated measure ANOVA comparing the means. Bold font highlights parameters with significant differences. Significance levels are indicated by an asterisk where p < 0.05. Units: TP mg P L
1; NH4: mg N L
1; TOC mg C L
1. mean rain event

TP N-NH4 TOC cells aggregates Tot ARG

41.6 97 0.844 789.4 29.2 6584

±5.46 ±21.14 ±0.14 ±134 ±8 ±3521

mean year

26.00 49.33 0.68 551.17 12.17 762

one way RM ANOVA

±8.06 ±32.2 ±0.24 ±246 ±4.10 ±857

3.3. Direct and indirect influence of rainfall on ARGs, bacterial and chemical parameters Table 3 shows that TP, NH4, aggregate number, all relative gene abundances and the total absolute ARG abundance were significantly influenced by the amount of rain that had fallen in the preceding three hours (P-value < 0.05, model 1). On the contrary, rainfall did not significantly influence TOC and single cell numbers (P-value > 0.05). Consequently, we tested whether the total absolute ARG numbers were influenced by the rain over the changes in TP and NH4 (Table 3, model 2) or aggregate numbers (Table 3, model 3). However, neither parameter yield a significant relationship with total ARGs (P-value > 0.05). 4. Discussion The temporal and spatial dynamics of ARGs in freshwater have only recently started to be addressed (Czekalski et al., 2015, 2016; Di Cesare et al., 2015) and further studies are needed to fully understand the source (hospitals, farms, industries, urban WWTPs; He et al., 2016) and the cycle of ARB and ARGs in the aquatic environment. Therein our study is the first to consider the impact of a moderate rain event on the dynamics of ARGs in freshwaters like a large but only slightly urbanized river. Until now, the influence of rainfalls on the abundance of ARGs has only been hypothesized (Novo et al., 2013), indirectly estimated by class 1 integrons prevalence (int1 gene copies per 16S rDNA copies) in River Thames (Amos et al., 2015), or tested by cultural approach in two lakes by analyzing surface water and surface sediment samples, before and after a storm water event (Zhang et al., 2016). All these studies suggest a potential impact of rainfalls on the microbial resistome of a water body in highly urbanized areas. We measured ARG abundances and biotic and abiotic variables in the River Toce monthly throughout the course of a year and specifically during a rain event (47.2 mm of rain over 2 days) that occurred after an exceptionally long dry period. Compared to the general rainfall in the Toce catchment area in 2016 that reached values of up to 130.8 mm in 2.5 days at the beginning of June, the rain event studied here can be rated as “moderate”. In our study, a significant increase of the absolute abundances of

Table 3 Formula, F- and P-values of the three tested linear models (model 1e3). Parameters with significant models are marked in bold font. Significance levels are indicated by asterisks where p < 0.05 ¼ *, p < 0.01 ¼ **. Model 1: lm(X ~ rain) X

F value

P value

Phosphorous N-NH4 TOC cells aggregates sulII/16S tetA/16S ermB/16S qnrS/16S Total ABR

10.99 32.44 2.67 1.732 10.94 21.98 8.489 17.49 18.46 8.503

0.0295* 0.0047** 0.177 0.258 0.0297* 0.009** 0.0435* 0.0139* 0.0127* 0.0434*

7.028 1.648

0.0769 0.2895

1.143

0.345

F-value

P-value

Model 2: lm(total ABR ~ TP þ NH4)

5.37 4.721 2.498 2.518 13.575 10.39

0.081 0.044* 0.189 0.188 0.021* 0.032*

TP NH4 Model 3: lm(total ABR ~ Aggregates) aggregates

Please cite this article in press as: Di Cesare, A., et al., Rainfall increases the abundance of antibiotic resistance genes within a riverine microbial community, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.04.036

A. Di Cesare et al. / Environmental Pollution xxx (2017) 1e6

all the tested ARGs during the rainfall was observed, which were strongly correlated with each other (Fig. 1; Tables 1 and 3). Here we limited the selection of genes to be quantified to those that were known to be abundant in the area (Di Cesare et al., 2015,2016b,c), and we cannot predict whether other ARGs would follow similar patterns. However, considering that tetA and sulII are typically harbored in gram-negative bacteria, qnrS can also be found in gram-positive (Rodríguez-Martínez et al., 2008), and ermB is a common ARG in gram-positive bacteria, our data suggest an unspecific and general response of the river resistome to the modified conditions during the rain. This resulted in a higher relative proportion of potential antibiotic resistant bacteria (expressed as ARG 16 S
1) which suggests a general shift of the microbial community composition as a consequence of the rainfall. After three weeks, the ARGs composition of the microbial community had returned to a pre-rain event state. Further investigations are needed to determine the later fate of the ARGs after entering the river, and which factors favour the observed resilience of the resident microbial community. This study also highlights a significant relation between the total and the relative ARG abundance, TP, N-NH4 and the number of microbial aggregates with the intensity of precipitation (Table 2). Microbial aggregation is of particular interest since it is commonly a response to stressors such as predation, antibiotics, UV or disinfectants (Corno and Jürgens, 2008; Corno et al., 2014; Di Cesare et al., 2016c; Modenutti et al., 2010). Moreover, aggregates are well recognized hot-spots for horizontal gene transfer (HGT), thus their presence potentially enhances the spread of ARGs within the microbial community (Barlow, 2009; Thomas and Nielsen, 2005). Here we did not find a direct statistical dependence of the ARG numbers to the numbers of aggregates. The relationship between the two might, however, not be linear, since aggregates are composed of many cells and can have a variety of shapes and sizes (Corno and Jürgens, 2008). HGT can be promoted under altered conditions and when different bacterial populations get in contact in spatially limited environments (Costerton et al., 1999), or when the main ecological drivers (competition, predation) are instable (Niehus et al., 2015). In comparison to aquatic environments, soils are generally inhabited by larger and more diverse bacterial communities, and their chemical and physical characteristics provide bacteria with a good substrate for the formation of multispecies biofilms where antibiotic resistance can be easily transfer from strain to strain by HGT (summarized by Ren et al., 2015). In fact, as consequence of disturbances as rainfalls, the soil biofilms can be dislodged, fractionated in aggregates, and suspended in surface waters (Stoodley et al., 2001). Furthermore, soils contain N and P compounds, which can be easily released into rivers and lakes during rainfalls (Bouraima et al., 2016), modifying the established nutrients ratio in river water. We found an increase of N-NH4 and TP concentrations during the rain event, as a result of runoff entering the river and of soil resuspension, with a consequent release of previously bounded P and N. In particular, stormwater originating during the event may have delivered nutrients deriving from residential or agricultural areas to the river (Hathaway et al., 2012). Altogether, our results support the hypothesis that the ARGs, nutrients, and aggregates found in the Toce water during the rain event derive from the catchment. Even though the sampled rain event was only a moderate, ARG absolute and relative abundances significantly exceeded the yearly average (Table 2). Moreover, the catchment area of the River Toce is rather poorly populated (<40 inhab. Km
2), and agriculture and stock-rearing activities are rather limited. Considering that many waters are characterized by more intense agriculture or

5

anthropogenic activities in the catchment area, rain events might lead to an even greater increase of ARGs in other rivers. Furthermore, prolonged dry periods before the rain event could result in the accumulation and following concomitant delivering of ARGs, nutrients and microbial aggregates to river water. 5. Conclusions This study highlights the impact of moderate rainfalls on the increase of potential antibiotic resistant bacteria and genes within aquatic microbial communities, suggesting the catchment as a potential source for allochthonous determinants of resistance. Consequently, disturbances related to meteorological conditions such as rainfalls should be considered within the monitoring on ARB/ARGs in aquatic environments in order to fully understand the routes of resistances spread and to take correct measures to reduce their abundance in natural environments. Contributors ADC, EME and GC designed the experiment. ADC and GC conducted the sampling, DNA extractions, qPCRs and gel electrophoresis. EME made statistics, tables, and graphs. MR measured chemical variables. All the authors analyzed and discussed the data, and wrote the paper. Acknowledgements We thank Cristiana Callieri for the productive discussions on the data, Mario Contesini and the staff of the CNR-ISE Hydrochemical Laboratory for the chemical analyses. This research is supported by the International Commission for the Protection of Italian-Swiss Waters (CIPAIS), Research Campaign 2016e18. Precipitation data were provided by the Regional Agency for Environmental Protection of Piedmont Region (ARPA Piemonte). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.envpol.2017.04.036. References American Meteorological Society, 2000. In: Cairns, M.M. (Ed.), Glossary of Meteorology. Amos, G.C., Zhang, L., Hawkey, P.M., Gaze, W.H., Wellington, E.M., 2014. Functional metagenomic analysis reveals rivers are a reservoir for diverse antibiotic resistance genes. Vet. Microbiol. 171, 441e447. Amos, G.C.A., Gozzard, E., Carter, C.E., Mead, A., Bowes, M.J., Hawkey, P.M., Zhang, L., Singer, A.C., Gaze, W.H., Wellington, E.M.H., 2015. Validated predictive modelling of the environmental resistome. ISME J. 9 (6), 1467e1476. APHA, AWWA, WEF, 2012. Standard Methods for Examination of Water and Wastewater. American Public Health Association, Washington. Barlow, M., 2009. What antimicrobial resistance has taught us about horizontal gene transfer. Methods Mol. Biol. 532, 397e411. Berendonk, T.U., Manaia, C.M., Merlin, C., Fatta-Kassinos, D., Cytryn, E., Walsh, F., €m, M., Pons, M.-N., Kreuzinger, N., Bürgmann, H., Sørum, H., Norstro Huovinen, P., Stefani, S., Schwartz, T., Kisand, V., Baquero, F., Martinez, J.L., 2015. Tackling antibiotic resistance: the environmental framework. Nat. Rev. Microbiol. 13, 310e317. Bouraima, A.K., He, B., Tian, T., 2016. Runoff, nitrogen (N) and phosphorus (P) losses from purple slope cropland soil under rating fertilization in Three Gorges Region. Environ. Sci. Pollut. Res. Int. 23 (5), 4541e4550. Bustin, S.A., Benes, V., Garson, J.A., Hellemans, J., Huggett, J., Kubista, M., Mueller, R., Nolan, T., Pfaffl, M.W., Shipley, G.L., Vandesompele, J., Wittwer, C.T., 2009. The MIQE guidelines - minimum information for publication of quantitative realtime PCR experiments. Clin. Chem. 55 (4), 611e622. Chen, D.J., Hu, M.P., Wang, J.H., Guo, Y., Dahlgren, R.A., 2016. Factors controlling phosphorus export from agricultural/forest and residential systems to rivers in eastern China, 1980-2011. J. Hydrology 533, 53e61. Corno, G., Jürgens, K., 2008. Structural and functional patterns of bacterial

Please cite this article in press as: Di Cesare, A., et al., Rainfall increases the abundance of antibiotic resistance genes within a riverine microbial community, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.04.036

6

A. Di Cesare et al. / Environmental Pollution xxx (2017) 1e6

communities in response to protist predation along an experimental productivity gradient. Environ. Microbiol. 10 (10), 2857e2871. Corno, G., Villiger, J., Pernthaler, J., 2013. Coaggregation in a microbial predator-prey system affects competition and trophic transfer efficiency. Ecology 94 (4), 870e881. Corno, G., Coci, M., Giardina, M., Plechuk, S., Campanile, F., Stefani, S., 2014. Antibiotics promote aggregation within aquatic bacterial communities. Front. Microbiol. 5, 297. Costerton, J.W., Stewart, P.S., Greenberg, E.P., 1999. Bacterial biofilms: a common cause of persistent infections. Science 284 (5418), 1318e1322. Cytryn, E., 2013. The soil resistome: the anthropogenic, the native, and the unknown. Soil Biol. Biochem. 63, 18e23. Czekalski, N., Berthold, T., Caucci, S., Egli, A., Bürgmann, H., 2012. Increased levels of multiresistant bacteria and resistance genes after wastewater treatment and their dissemination into Lake Geneva. Switz. Front. Microbiol. 3, 106. Czekalski, N., Diez, E., Burgmann, H., 2014. Wastewater as a point source of antibiotic resistance genes in the sediment of a freshwater lake. ISME J. 8, 1381e1390. Czekalski, N., Sigdel, R., Birtel, J., Matthews, B., Bürgmann, H., 2015. Does human activity impact the natural antibiotic resistance background? Abundance of antibiotic resistance genes in 21 Swiss lakes. Environ. Int. 81, 45e55. Czekalski, N., Imminger, S., Salhi, E., Veljkovic, M., Kleffel, K., Drissner, D., Hammes, F., Bürgmann, H., von Gunten, U., 2016. Inactivation of antibiotic resistant bacteria and resistance genes by ozone: from laboratory experiments to full-scale wastewater treatment. Environ. Sci. Technol. 50 (21), 11862e11871. Di Cesare, A., Luna, G.M., Vignaroli, C., Pasquaroli, S., Tota, S., Paroncini, P., Biavasco, F., 2013. Aquaculture can promote the presence and spread of antibiotic-resistant enterococci in marine sediments. PLoS One 26 (8e4), e62838. Di Cesare, A., Eckert, E.M., Teruggi, A., Fontaneto, D., Bertoni, R., Callieri, C., Corno, G., 2015. Constitutive presence of antibiotic resistance genes within the bacterial community of a large subalpine lake. Mol. Ecol. 24, 3888e3900. Di Cesare, A., Eckert, E.M., Corno, G., 2016a. Co-selection of antibiotic and heavy metal resistance in freshwater bacteria. J. Limnol. 75 (s2), 59e66. Di Cesare, A., Eckert, E.M., D'Urso, S., Bertoni, R., Gillan, D.C., Wattiez, R., Corno, G., 2016b. Co-occurrence of integrase 1, antibiotic and heavy metal resistance genes in municipal wastewater treatment plants. Water Res. 94, 208e214. Di Cesare, A., Fontaneto, D., Doppelbauer, J., Corno, G., 2016c. Fitness and recovery of bacterial communities and antibiotic resistance genes in urban wastewaters exposed to classical disinfection treatments. Environ. Sci. Technol. 50 (18), 10153e10161. Eckert, E.M., Di Cesare, A., Stenzel, B., Fontaneto, D., Corno, G., 2016. Daphnia as a refuge for an antibiotic resistance gene in an experimental freshwater community. Sci. Total Environ. 571, 77e81. Hathaway, J.M., Tucker, R.S., Spooner, J.M., Hunt, W.F., 2012. A traditional analysis of the first flush effect for nutrients in stormwater runoff from two small urban catchments. Water Air Soil Pollut. 223, 5903e5915. Hatosy, S.M., Martiny, A.C., 2015. The ocean as a global reservoir of antibiotic resistance genes. Appl. Environ. Microbiol. 81, 7593e7599. He, L.-Y., Ying, G.-G., Liu, Y.-S., Su, H.-C., Chen, J., Liu, S.-S., Zhao, J.-L., 2016. Discharge of swine wastes riskswater quality and food safety: antibiotics and antibiotic resistance genes from swine sources to the receiving environments. Environ. Int. 92e93, 210e219.
a k, K., Corno, G., 2012. Every coin has a back side: invasion by Limnohabitans Horn planktonicus promotes the maintenance of species diversity in bacterial communities. PLoS One 7 (12), e51576. Janke, B.D., Finlay, J.C., Hobbie, S.E., Baker, L.A., Sterner, R.W., Nidzgorski, D., Wilson, B.N., 2014. Contrasting influences of stormflow and baseflow pathways on nitrogen and phosphorus export from an urban watershed. Biogeochemistry 121, 209e228. Ju, F., Li, B., Ma, L., Wang, Y., Huang, D., Zhang, T., 2016. Antibiotic resistance genes and human bacterial pathogens: Co-occurrence, removal, and enrichment in municipal sewage sludge digesters. Water Res. 91, 1e10.

Lee, M.-H., Payeur-Poirier, J.-L., Park, J.-H., Matzner, E., 2016. Variability in runoff fluxes of dissolved and particulate carbon and nitrogen from two watersheds of different tree species during intense storm events. Biogeosciences 13, 5421e5432. Modenutti, B., Balseiro, E., Corno, G., Callieri, C., Bertoni, R., Caravati, E., 2010. Ultraviolet radiation induces filamentation in bacterial assemblages from North Andean Patagonian lakes. Photochem. Photobiol. 86 (4), 871e881. National Research Council of Italy - Institute of Ecosystem Study, 2015. In: Commissione Internazionale per la protezione delle acque italo-svizzere (Ed.), Ricerche sull'evoluzione del Lago Maggiore. Aspetti limnologici. Programma Triennale 2013-2015, p. 003103. Niehus, R., Mitri, S., Fletcher, A.G., Foster, K.R., 2015. Migration and horizontal gene transfer divide microbial genomes into multiple niches. Nat. Commun. 6, 8924. , S., Viana, P., Nunes, O.C., Manaia, C.M., 2013. Antibiotic resistance, Novo, A., Andre antimicrobial residues and bacterial community composition in urban wastewater. Water Res. 47 (5), 1875e1887. Outram, F.N., Lloyd, C.E.M., Jonczyk, J., Benskin, C.M.H., Grant, F., Perks, M.T., Deasy, C., Burke, S.P., Collins, A.L., Freer, J., Haygarth, P.M., Hiscock, K.M., Johnes, P.J., Lovett, A.L., 2014. High-frequency monitoring of nitrogen and phosphorus response in three rural catchments to the end of the 2011e2012 drought in England. Hydrology Earth Syst. Sci. 18, 3429e3448. Popowska, M., Rzeczycka, M., Miernik, A., Krawczyk-Balska, A., Walsh, F., Duffy, B., 2012. Influence of soil use on prevalence of tetracycline, streptomycin, and erythromycin resistance and associated resistance genes. Antimicrob. Agents Chemother. 56, 1434e1443. Pruden, A., Pei, R., Storteboom, H., Carlson, K.H., 2006. Antibiotic resistance genes as emerging contaminants: studies in Northern Colorado. Environ. Sci. Technol. 40 (23), 7445. Pruden, A., Arabi, M., Storteboom, H.N., 2012. Correlation between upstream human activities and riverine antibiotic resistance genes. Environ. Sci. Technol. 46, 11541e11549. R Core Team, 2013. R: a Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. RStudio Team, 2015. RStudio: Integrated Development for R. RStudio, Inc., Boston, USA. Ravazzani, G., Dalla Valle, F., Gaudard, L., Mendlik, T., Gobiet, A., Mancini, M., 2016. Assessing climate impacts on hydropower production: the case of the Toce river basin. Climate 4, 16. Ren, D., Madsen, J.S., Sørensen, S.J., Burmølle, M., 2015. High prevalence of biofilm synergy among bacterial soil isolates in cocultures indicates bacterial interspecific cooperation. ISME J. 9 (1), 81e89. Rodríguez-Martínez, J.M., Velasco, C., Briales, A., García, I., Conejo, M.C., Pascual, A., 2008. Qnr-like pentapeptide repeat proteins in gram-positive bacteria. J. Antimicrob. Chemother. 61 (6), 1240e1243. Schwartz, T., 2012. In: Keen, P.L., Montforts, M.H.M.M. (Eds.), Antimicrobial Resistance in the Environment. Wiley-Blackwell, pp. 251e264. Stoodley, P., Wilson, S., Hall-Stoodley, L., Boyle, J.D., Lappin-Scott, H.M., Costerton, J.W., 2001. Growth and detachment of cell clusters from mature mixed-species biofilms. Appl. Environ. Microbiol. 67 (12), 5608e5613. Taylor, N.G.H., Verner-Jeffreys, D.W., Baker-Austin, C., 2011. Aquatic systems: maintaining, mixing and mobilising antimicrobial resistance? Trends Ecol. Evol. 26, 278e284. Thomas, C.M., Nielsen, K.M., 2005. Mechanisms of, and barriers to, horizontal gene transfer between bacteria. Nat. Rev. Microbiol. 3, 711e721. Zhang, S., Pang, S., Wang, P., Wang, C., Han, N., Liu, B., Han, B., Li, Y., Anim-Larbi, K., 2016. Antibiotic concentration and antibiotic-resistant bacteria in two shallow urban lakes after stormwater event. Environ. Sci. Pollut. Res. Int. 23 (10), 9984e9992. Zurfuh, K., Poirel, L., Nordmann, P., Nüesch-Inderbinen, M., H€ achler, H., Stephan, R., 2016. Occurrence of the plasmid-borne mcr-1 colistin resistance gene in ESBLproducing Enterobacteriacae in river water and imported vegetable samples in Switzerland. Antimicrob. Agents Chemother. 60 (4), 2594e2595.

Please cite this article in press as: Di Cesare, A., et al., Rainfall increases the abundance of antibiotic resistance genes within a riverine microbial community, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.04.036