Effects of Ciprofloxacin Alone or in Mixture with Sulfamethoxazole on the Efficiency of Anaerobic Digestion and Its Microbial Community

Some livestock farms rely on anaerobic digestion (AD) technology for manure disposal, thus obtaining energy (biogas) and fertilizer (digestate). Mixtures of antibiotics used for animal health often occur in organic waste and their possible synergistic/antagonistic effects on microorganisms involved in AD are still poorly studied. This work focuses on the effects of adding ciprofloxacin, alone (5 mg L−1) and in combination with sulfamethoxazole (2.5–5–10 mg L−1), on AD efficiency and microbial community structure. The experiment consisted of 90-day cattle manure batch tests and antibiotic removal percentages were assessed. Adding antibiotics always promoted CH4 and H2 production compared to untreated controls; however, CH4 production was lowered with the highest ciprofloxacin (CIP) concentrations. The overall results show antibiotic degradation caused by acidogenic Bacteria, and CH4 was mainly produced through the hydrogenotrophic-pathway by methanogenic Archaea. Shifts in microbial community abundance (DAPI counts) and composition (Illumina-MiSeq and FISH analyses) were observed.


Introduction
An increasing number of intensive livestock farms are relying on anaerobic digestion (AD) technology to solve the disposing of the huge amount of manure and slurry produced daily [1], getting the benefit of energy and fertilizer from biogas and digestate, respectively [2]. The biogas obtained, generally containing 50-75% CH 4 , can be burned in co-generation engines to obtain electric energy [3] or purified to biomethane [4]. The digestate remaining after AD treatment is commonly spread on the soil as a replacement for conventional mineral fertilizers. In the last decade, AD treatment has also been attracting interest for its potential to degrade some organic pollutants and emerging contaminants like pharmaceuticals commonly found in animal manure [5,6], although much research is needed to verify the effectiveness of the process. Veterinary antibiotics are widely used to treat and prevent animal diseases in husbandry practices. Despite their undoubted benefits, their massive use is leading to several risks for environmental processes and ecosystem services [7,8]. Moreover, inhibition of engineered bioprocesses for slurry disposal, such as AD [1], is also suspected. Animals excrete the largest amount (between 10% and 90%) of the Figure 1. Cumulative H2 and CH4 production in all experimental conditions. Dotted lines and solid lines represent the cumulative production of H2 and CH4, respectively (standard deviations < 10%). The values of the SMX_5 test are those reported in Mazzurco Miritana et al. [33].
In all cases, H2 production started immediately, reaching the maximum value at days 6-8. The start of CH4 production corresponded to the H2 decrease in the biogas on the timelines ( Figure 1). The cumulative amount of H2 detected was comparable to that of CH4 production, for all experimental conditions (values ranging from 753.7 ± 102.4 mL L −1 to 2040.3 ± 20.2 mL L −1 detected for control and MIX_5, respectively). The control condition showed the lowest cumulative values of both H2 (p < 0.01) and CH4 (p < 0.001) and was also the only case showing twice as much H2 as CH4 (899.5 ± 101.1 mL L −1 and 386.8 ± 65.4 mL L −1 , respectively).
The highest CH4 cumulative production (p < 0.01) was detected in MIX_2.5 with a value more than six-fold higher (2.742 ± 369 mL L −1 ) than the control. No significant differences were found at the end of the experiment among the other three spiked conditions, i.e., CIP_5, MIX_5 and MIX_10 (1.985 ± 228 mL L −1 , 1.680 ± 148 mL L −1 and 1.919 ± 136 mL L −1 , respectively). Despite this, the condition with only one antibiotic, i.e., CIP_5, reached maximum CH4 production already on day 40 of the experiment, whereas MIX_5 and MIX_10 needed more time, achieving the maximum cumulative production at days 68 and 83, respectively. Interestingly, among the conditions with the same amount of added antibiotics, i.e., MIX_2.5 and CIP_5, the latter, containing only ciprofloxacin, produced less CH4, revealing a detrimental effect if compared to its action in combination with SMX. A study carried out in the same experimental condition and spiking SMX alone at a concentration of 5 mg L −1 [33], achieved a final cumulative CH4 production of 2030.6 ± 143.3 mL L −1 , a value comparable with those obtained in this experiment (excluding control and MIX_2.5 tests). Nevertheless, the authors observed that CH4 production in the SMX spiked test ended after only 33 days. This result confirms a lower impact of SMX on the AD microbial community than CIP, also due to its higher degradability. Other interesting information was obtained from the results on the composition (%) of the biogas at the end of the experiment ( Figure S1). The control condition showed the lowest CH4 content in the biogas, with a concentration never exceeding 31%. Conversely, both CIP_5 and MIX_5 showed more than doubled CH4 concentrations, 73.1 ± 8.4% and 68.8 ± 6.4%, respectively, although these values were reached at days 35 and 69, respectively. In MIX_10 the CH4 content was 58.5 ± 6.4% (day 69). The highest CH 4 cumulative production (p < 0.01) was detected in MIX_2.5 with a value more than six-fold higher (2.742 ± 369 mL L −1 ) than the control. No significant differences were found at the end of the experiment among the other three spiked conditions, i.e., CIP_5, MIX_5 and MIX_10 (1.985 ± 228 mL L −1 , 1.680 ± 148 mL L −1 and 1.919 ± 136 mL L −1 , respectively). Despite this, the condition with only one antibiotic, i.e., CIP_5, reached maximum CH 4 production already on day 40 of the experiment, whereas MIX_5 and MIX_10 needed more time, achieving the maximum cumulative production at days 68 and 83, respectively. Interestingly, among the conditions with the same amount of added antibiotics, i.e., MIX_2.5 and CIP_5, the latter, containing only ciprofloxacin, produced less CH 4 , revealing a detrimental effect if compared to its action in combination with SMX. A study carried out in the same experimental condition and spiking SMX alone at a concentration of 5 mg L −1 [33], achieved a final cumulative CH 4 production of 2030.6 ± 143.3 mL L −1 , a value comparable with those obtained in this experiment (excluding control and MIX_2.5 tests). Nevertheless, the authors observed that CH 4 production in the SMX spiked test ended after only 33 days. This result confirms a lower impact of SMX on the AD microbial community than CIP, also due to its higher degradability. Other interesting information was obtained from the results on the composition (%) of the biogas at the end of the experiment ( Figure S1). The control condition showed the lowest CH 4 content in the biogas, with a concentration never exceeding 31%. Conversely, both CIP_5 and MIX_5 showed more than doubled CH 4 concentrations, 73.1 ± 8.4% and 68.8 ± 6.4%, respectively, although these values were reached at days 35 and 69, respectively. In MIX_10 the CH 4 content was 58.5 ± 6.4% (day 69).
A wide range of effects from complete inhibition to stimulation of CH 4 production, depending on the antibiotic types and concentrations, was previously observed by other authors [7,9,27]. The aforementioned study by Mazzurco Miritana and colleagues [33] showed that the addition of SMX (5 mg L −1 ) to cattle manure resulted in higher CH 4 production compared to the control condition. A similar result was reported by Zhi and Zhang [27] who observed that the addition of 100 mg L −1 of CIP to cattle manure stimulated CH 4 production. Even in this study, the addition of antibiotics stimulated CH 4, as well as H 2 production, highlighting that these results need further investigation to be fully understood. In particular, it needs to be verified whether antibiotics, in this case CIP, exerted selective pressure on specific functional components of the microbial community, probably competing with methanogens for the use of H 2 (i.e., bacteria performing sulphate reduction and/or homoacetogenesis).
From this study emerges that the mixed conditions had the highest H 2 production, and the latter hindered the AD process by delaying the start of CH 4 production. Typically, the H 2 produced during the AD process is not detected in significant concentrations in the biogas (i.e.,~5%) since it is mainly consumed by the microbial community for the final production of CH 4 . In this study, the high detection of H 2 proved to be an indicator of the imbalance of the different phases of the AD process as widely discussed in Mazzurco Miritana et al. [33]. In particular, the H 2 produced by Bacteria during the hydrolytic/acidogenic and acetogenic AD phases was not used by Archaea to produce CH 4 as quickly as it was produced. Some studies reported that adding antibiotics inhibited acetate, propionate, and butyrate uptake [30,34]. Another study [35] reported possible effects of fluoroquinolones and sulfonamides on the Archaea domain affecting microbial growth and activity. In this study, the imbalance of the metabolic phases of the AD resulted in a process configured in two phases, with an earlier H 2 production followed by a CH 4 production phase. A similar condition was already observed in a previous study [36] in which imbalances between different functional groups in the microbial community resulted in imbalances in the AD process. The interpretation of the process imbalance is also supported by the results obtained from the analysis of the VFAs, lactic acid and ethanol, which were intermediates of the process (Figure 2a-e).
Although the ingestate collected already contained high concentrations of acetic (3.7 g L −1 ), propionic (2.5 g L −1 ) and lactic acids (1.6 g L −1 ) as well as ethanol (4.3 g L −1 ), in all cases the H 2 production started immediately at the beginning of the experiment. Indeed, from the first two experimental weeks, a very high accumulation of acetic acid was observed in all experimental conditions, including the control (Figure 2a). The values even above 10,000 mg L −1 indicated the inability of the microbial community to promptly use it for producing methane. A previous study [4] reported that elevated hydrogen levels could temporarily inhibit the AD process by promoting the accumulation of acetate. In all spiked conditions, succinic, formic, valeric and butyric acids as well as ethanol were detected at values of a quarter and a fifth of the acetic acid (around 2000 mg L −1 ), indicating they were partially consumed during the acetogenic metabolic step (Figure 2b-e). An exception was represented by MIX_10, in which propionic acid persisted from day 9, with values around 6000 mg L −1 . In the control condition, the acid concentrations were higher than in the spiked tests, particularly for butyric, propionic and valeric ones. The detection of increasing concentrations of acetic acid revealed that the rapid production of acids, as well as their transformation into acetic acid, was not followed by their use for methane production through the acetoclastic pathway, highlighting that there was a bottleneck in the AD process between the acetogenic and methanogenic metabolic phases. In contrast, hydrogen consumption, in line with the start of CH 4 production, clearly showed that methanogenesis was performed by the hydrogenotrophic metabolic pathway.

Antibiotic Removal
In the course of the experiment, CIP and SMX were monitored, and the efficiency of their removal was quantified. The results obtained ( Figure 3) show that both CIP and SMX were degraded during the AD process. On the other hand, the antibiotic mixtures and increasing concentrations hindered their complete depletion.  genesis was performed by the hydrogenotrophic metabolic pathway.

Antibiotic Removal
In the course of the experiment, CIP and SMX were monitored, and the efficiency of their removal was quantified. The results obtained ( Figure 3) show that both CIP and SMX were degraded during the AD process. On the other hand, the antibiotic mixtures and increasing concentrations hindered their complete depletion. CIP alone (i.e., CIP_5) degraded three times less when combined with SMX in MIX_5, showing 20.9% and 6.5% removal, respectively. This difference, observed in tests conducted with the same final antibiotic concentration, is attributable to the synergism generated by the different mechanisms of action exerted by the two antibiotics investigated CIP alone (i.e., CIP_5) degraded three times less when combined with SMX in MIX_5, showing 20.9% and 6.5% removal, respectively. This difference, observed in tests conducted with the same final antibiotic concentration, is attributable to the synergism generated by the different mechanisms of action exerted by the two antibiotics investigated on microorganisms. The lowest and highest concentrations of antibiotic mixtures, i.e., MIX_2.5 and MIX_10, showed the highest (8.1%) and lowest (5.6%) CIP removal, respectively. In any case, CIP removal never exceeded 21%, confirming this antibiotic to be a recalcitrant compound in natural and artificial ecosystems, as reported by other authors [15,18,25]. SMX was degraded more easily than CIP. A percentage of 49.0% of its removal was reached in MIX_2.5, but its degradation was also strongly negatively influenced by the co-presence of CIP. Indeed, Mazzurco Miritana et al. [33] observed that SMX alone was removed 81% at the end of their experiment. Some aspects remain to be clarified concerning the role of SMX in the mixtures. Indeed, at the end of the experiment, Mix_10 produced more CH 4 than Mix_5 (Figure 1), in line with the removal rates ( Figure 3). On the other hand, SMX_5 produced CH 4 faster than CIP_5, since the production plateau was reached at days 22 (SMX_5) and 40 (CIP_5), showing a lower effect on the microbial community of SMX than that of CIP.
It is very interesting to note that, in all the experimental conditions, antibiotic removal only occurred until day 21, during H 2 production, corresponding to the hydrolytic/acidogenic metabolic step of AD. Indeed, the maximum H 2 peaks were reached within the first 21 days of the experiment, suggesting that the antibiotic removal was therefore ascribable to the hydrolytic-fermenting microorganisms able to break down complex molecules into simpler ones. On the contrary, Mazzurco Miritana et al. [33] observed that SMX alone degraded even after day 21, and at the end of the experiment (69 days) there was just 20% of its initial concentration.

Microbial Community Analysis
The values of the total microbial abundance were in line with the H 2 and CH 4 production trends. Under all spiked conditions, an initial detrimental effect of antibiotics on microbial abundance was found, as compared to the control (Figure 4a), demonstrating that antibiotics negatively affected microbial populations within the reactors. However, this effect was transient and an increase in microbial abundance was observed in the spiked conditions. The mixtures MIX_2.5, MIX_5 and MIX_10 showed the highest microbial abundances in the first 15 experimental days (days 15 and 6, respectively), in line with their maximum H 2 production rate ( Figure 1). In particular, MIX_2.5, the second one in cumulative H 2 production, and the first one in cumulative CH 4 production, showed the highest microbial abundances, at days 6 and 15 (6.2 × 10 9 ± 4.6 × 10 8 cells mL −1 and 7.91 × 10 9 ± 8.35 × 10 8 cells mL −1 , respectively), while MIX_5 reached its maximum microbial abundance at day 6 (7.5 × 10 9 ± 6.62 × 10 8 cells mL −1 ), in line with the highest H 2 production detected during the experiment. Conversely, the control condition showed the highest microbial abundance in the first experimental week and then gradually decreased. At the end of CH 4 production, microbial abundance values dropped in all experimental conditions.  A number of previous studies reported the effects of antibiotics on the Bacteria and Archaea domains [26,27,37], showing that macrolides inhibited more methanogens than Bacteria, in particular acetoclastic. Moreover, it must be considered that the Archaea component of the AD microbial community thrives by relying on Bacteria metabolism. The antibiotic impact on specific functional bacterial components could also therefore result in the indirect and unexpected elimination of antibiotic-resistant Archaea [38]. This study shows that the highest CH4 production efficiency observed under spiked conditions The FISH analysis clearly showed that adding antibiotics always had detrimental effects on archaeal abundance (N. cells mL −1 ) (Figure 4b) as well as on Archaea percentages (%) (Figure 4c), in line with antibiotic concentrations. A number of previous studies reported the effects of antibiotics on the Bacteria and Archaea domains [26,27,37], showing that macrolides inhibited more methanogens than Bacteria, in particular acetoclastic. Moreover, it must be considered that the Archaea component of the AD microbial community thrives by relying on Bacteria metabolism. The antibiotic impact on specific functional bacterial components could also therefore result in the indirect and unexpected elimination of antibiotic-resistant Archaea [38]. This study shows that the highest CH 4 production efficiency observed under spiked conditions should not be sought solely in the composition and abundance of microbial components, but further studies are needed to identify the AD parameters affecting microbial community functionality.
A few studies have discussed the proportion of fermentative Bacteria versus methanogenic Archaea [36,39], although it could prove to be a key factor in assessing the efficiency of the AD process. To our knowledge, this study reported the proportion of Bacteria/Archaea cells under different experimental conditions and at different experimental times for the first time (Table 1). It is made evident that the control condition showed few variations during the experiment while the spiked conditions, particularly the MIXs, showed the highest values, thus demonstrating an increased proportion of Bacteria. A similar result was reported by Mustapha et al. [39]. The authors reported that azithromycin in sludge promoted an increased proportion of Bacteria, corresponding, in this case, to an increase in hydrolysis efficiency. In turn, a high bacterial concentration can lead to a further accumulation of the intermediate VFAs (if they are not readily converted to methane), as observed also in this study. Moreover, although Archaea can be resistant to antibiotics, some metabolic groups appear to be more susceptible [38]. A previous study reported that hydrogenotrophic methanogens had low sensitivity to CIP and SMX antibiotics [26].
The microbial community analysis performed using the MiSeq Illumina platform (Figure 5a,b) showed initial percentages of 99% Bacteria and <1% Archaea, respectively (Figure 5b). Methanobrevibacter and Methanosphaera were the dominant genera in the Archaea guild (Figure 5a), with the latter dominant at the end of the experiment in the MIXs conditions. Both genera produce methane through the hydrogenotrophic pathway. In particular, Methanosphera obtain energy for growth by using hydrogen to reduce methanol to methane [40]. The results confirm that hydrogenotrophic methanogens predominated in the microbial community, as was supposed on the basis of the results obtained from the analysis of process intermediates and trends in H 2 content in the biogas. Only a small fraction of acetoclastic genera, Methanosarcina and Methanosaeta, were detected. Both genera produce methane through the hydrogenotrophic pathway. In particular, Methanosphera obtain energy for growth by using hydrogen to reduce methanol to methane [40]. The results confirm that hydrogenotrophic methanogens predominated in the microbial community, as was supposed on the basis of the results obtained from the analysis of process intermediates and trends in H2 content in the biogas. Only a small fraction of acetoclastic genera, Methanosarcina and Methanosaeta, were detected.
An analysis of the microbial community at class level ( Figure 5b) showed that Clostridia made up more than half of the community and only declined under mixed conditions at the end of the experiment. Actinobacteria, Bacilli, Bacteroidia and Gammaproteobacteria were the other main components of the community. An analysis of the microbial community at class level (Figure 5b) showed that Clostridia made up more than half of the community and only declined under mixed conditions at the end of the experiment. Actinobacteria, Bacilli, Bacteroidia and Gammaproteobacteria were the other main components of the community.
The principal coordinate analysis based on Bray-Curtis distances ( Figure 6) showed significant differences (Permanova, p < 0.001) among the conditions considering the prokaryotic communities.
The principal coordinate analysis based on Bray-Curtis distances ( Figure 6) showed significant differences (Permanova, p < 0.001) among the conditions considering the prokaryotic communities. The diversity indices are reported in Table 2. The Chao1 and Shannon (H) indices in antibiotics treated samples were lower than in the control at all sampling times. This decrease in diversity can be associated with the detrimental effect of the antibiotics on the prokaryotic community. Among the antibiotic-treated conditions, the MIXs at The diversity indices are reported in Table 2. The Chao1 and Shannon (H) indices in antibiotics treated samples were lower than in the control at all sampling times. This decrease in diversity can be associated with the detrimental effect of the antibiotics on the prokaryotic community. Among the antibiotic-treated conditions, the MIXs at 15 and 21 days showed microbial diversity values lower than CIP_5 (Table 2). This result can be ascribed to a stronger combined effect of the antibiotic mixtures on some microbial populations. However, the decrease in microbial diversity did not affect the community functionality since antibiotic spiked conditions produced higher amounts of CH 4 than the control, showing the presence of resistant prokaryotes. In particular, it can be stated that antibiotics exerted a selective effect on methanogenic Archaea favoring the genera of Methanobrevibacter and Methanosphera, which were very effective in the production of methane along the hydrogenotrophic metabolic pathway.

Anaerobic Digestion Test
The experimental setup of the AD tests is detailed in a previous work [33]. Briefly, the feeding ingestate of a full-scale AD plant, consisting of fresh cattle manure, was used to perform batch AD tests spiked with antibiotics. In particular, the ingestate was collected from the feeding tank pipe of a full-scale reactor located in a beef and dairy cattle farm in Central Italy (Lazio region). The tank collected the cattle manure produced daily and, using a pump, sent it to the reactor. Aliquots were collected by filling glass bottles of 2 L each (three replicates) after the pump of the collecting tank had been running for 15 min. The bottles were transported to the laboratory where the ingestate was analyzed to detect possible residual amounts of SMX (0.3 mg L −1 ) and CIP (0.1 mg L −1 ) before setting up the experiment. The bottles were transported to the laboratory where the experiment was immediately set up. The ingestate used was characterized as reported in a previous work [33]. Five experimental conditions (3 replicates each) were set up by adding antibiotics as follows: CIP alone (5 mg L −1 ), three mixtures of CIP and SMX (1:1 ratio) at concentrations of 2.5, 5, 10 mg L −1 of each antibiotic (hereinafter referred as CIP_5, MIX_2.5, MIX_5, MIX_10, respectively) and a control condition, set up without antibiotic spiking. The antibiotic concentrations were established by taking into account our previous study evaluating the presence of SMX, CIP and ENR in a full-scale anaerobic plant [17] located in the same geographical area.
Antibiotic single stock solutions were prepared by dissolving powdered CIP (purity 99%, Sigma-Aldrich) and SMX (purity 99%, Sigma-Aldrich) in a methanol/MilliQ solvent. The solutions obtained were further diluted in ultrapure water to reach the final concentrations to be tested.
The AD tests were set up using 600 mL Pyrex glass bottles filled with 300 mL of ingestate corresponding to 31 g volatile solids (VS). The bottles were sealed with a rubber stopper and metal ring and flushed with N 2 (10 s) to establish anaerobic conditions. Antibiotic solutions were added by using syringes and each bottle was shaken for 10 s. A solution of ultrapure water and methanol was added to the control conditions. The final amount of methanol in each batch was 0.18 mL. The samples were immediately collected in order to check the concentration of antibiotics at the beginning of the experiment. Throughout the experiment, the reactors were kept at a constant temperature of 37 to 38 • C. Measurements of biogas production and its CH 4 content, as well as liquid medium sampling, were performed daily during the first 6 experimental days. Subsequently, samples were collected once a week until day 27, then once every fortnight until the end of biogas production. Each experimental test was considered completed when no more CH 4 was produced for two weeks [36].

Biogas and Organic Acid Measurements
Biogas volumetric production was measured using a water displacement device [41] and H 2 , CH 4 and CO 2 content (%) in the biogas was analyzed using a gas chromatograph (Focus GC, by Thermo Scientific, Waltham, MA, USA) equipped with a thermal conductivity detector (TCD) and a 3 m stainless-steel column packed with Hayesep Q (800/100 mesh). Nitrogen gas was used as a carrier at a flow rate of 35 mL min −1 . The temperature of the column and of the injector was 120 • C, while that of the TCD was 200 • C.
Cumulative H 2 and CH 4 productions were calculated with Logan equation [42]. The composition of process intermediates in terms of acetic, succinic, lactic, formic, propionic, butyric and valeric acids, as well as ethanol, was analyzed using High Performance Liquid Chromatograph (HPLC) technology. Liquid samples were diluted 1:10 in H 2 SO 4 5 mN and filtered with a 0.22 µm membrane before injection into the HPLC. A Thermo Spectra system (USA), equipped with a UV detector (λ 210 nm) and 300 mm × 7.8 mm Rezex ROA-Organic Acid Hþ (8%) column (Phenomenex, Torrance, CA, USA) with a 4 × 30 mm Carbo-H security guard cartridge (Phenomenex, USA), was operated at 75 • C, by using a 5 mN H 2 SO 4 solution as the mobile phase (flow rate, 0.5 mL min −1 ).

Antibiotic Determination
The analytical determination of SMX and CIP were performed as reported by Visca et al. [17].  [17]. The HPLC-MS/MS system and data acquisition were controlled by the Analyst ® 1.6 Software (AB Sciex, Concord, ON, Canada).
A good linearity was obtained for the concentration range of 0.5-7.5 mg L −1 for the two antibiotics, as indicated by the correlation coefficient (R 2 , always >0.99). Calibration standards (in the range of 0.1-7.5 mg L −1 ) were prepared in triplicate for three validation runs performed on different days. The relative standard deviations of the concentration tested were <15%. The internal standard calibration was performed with the addition of deuterated standards (sulfamethoxazole-d4 and ciprofloxacin-d8 hydrochloride hydrate) to both working standard solutions and purified extracts. Recovery was evaluated by spiking feeding ingestate matrix with the target antibiotics alone and in a mixture (1:1 ratio) at three different concentrations (1, 5 and 10 mg L −1 , five replicates). The average recovery rates for SMX and CIP were 107.4 ± 6%, and 68.2 ± 5%, respectively. No reduction in the average recoveries was measured when antibiotics were spiked as a mixture.
The limits of detection (LOD) for the SMX and CIP antibiotics [43] were 1.5 µg L −1 and 2.0 µg L −1 , respectively. The quantification limits were set as three times LOD values.

Microbiological Analysis
Liquid medium samples (triplicates) were fixed as described in Pernthaler et al. [44] and stored (−20 • C) until use. The timing for microbiological investigations was chosen in line with the performance of the AD process. Total microbial abundance was determined at days 3, 6, 10, 15, 21, 27 and at the end of CH 4 production by using direct cell count (N. cells mL −1 ) after staining with DAPI (4 ,6-diamidino-2-phenylindole, 1 µg mL −1 each sample, 3 replicates, 20 ). The samples were then collected on black polycarbonate filters (pore size 0.22 µm, diameter 25 mm, Millipore, Burlington, MA, USA). Filters were placed on slides and examined (20 fields) using a Zeiss epifluorescence microscope AXIOSKOP 40 (Carl Zeiss, Jena, Germany) equipped with a ZEISS HXP 120v Light Source (1000× magnification). Microbial community structure was analyzed using the fluorescence in situ hybridization (FISH) technique as described in Amann et al. [45] and Barra Caracciolo et al. [46] at days 3, 6, 10, 15, 21 and at the end of CH 4 production. Before analysis, a cell extraction procedure was performed to detach and separate microbial cells from inorganic particles as described in Barra Caracciolo et al. [47]. Each sample (3 replicates) was collected on white polycarbonate filters (pore size 0.22 µm, diameter 25 mm, Millipore, Burlington, MA, USA). The EUB338, II, III and ARC915 probes were used for the detection of Bacteria and Archaea active cells [48]. The cell average enumeration was performed in order to evaluate the percentage of positive signals versus DAPI-stained positive cells (20 fields for each slide).

DNA Extraction and Next Generation Sequence (NGS)
The genome of the prokaryotic community was sequenced with a next-generation sequencing (NGS) using the MiSeq platform (Illumina). The Pro341F and Pro805R primers were used to amplify the V3-V4 region of 16S rRNA genes. These primers were selected in order to ensure the simultaneous identification of Bacteria and Archaea [49]. FastQ files were imported using QIIME2 v2019.11 [50] and denoised with the DADA2 plugin [51]. The primers were removed using the "trim-left-f" (forward) and "trim-left-r" (reverse) DADA2 functions. The length of the primers was 17 nucleotides for the forward and 21 nucleotides for the reverse one. The amplicon sequencing variants (ASVs) with less than 0.005% of high-quality reads were then filtered and compared to the 97% identical clustered Ribosomal Database Project (RDP release 11) using a naive Bayes classifier trained on the amplified region with an 80% confidence. The bacterial diversity was analyzed using the Evenness and Shannon diversity indices, while the Chao 1 index [52] was used as an estimator of potential richness (calculated with Qiime2 using the diversity alpha function).

Statistical Analysis
The Kruskal-Wallis test (a non-parametric one-way ANOVA) was used to evaluate the differences within the alpha diversity indices and was calculated by R (4.0.4 version https://www.r-project.org/ (accessed on 15 June 2022)), using the kruskal.test function together with the pairwise.wilcox.test function, as the post-hoc test [53].
The principal coordinate analysis based on the Bray-Curtis distance and PERMANOVA as the statistical test was performed using the online tool Microbiome Analyst (http://www. microbiomeanalyst.ca (accessed on 15 June 2022)) in order to evaluate the composition of the bacterial community during the AD process.
Finally, all the histograms and stacked bar plots were constructed with MS Excel 2013.

Conclusions
Adding SMX and CIP alone or in mixture did not negatively affect the AD process, showing that the functional prokaryotic populations involved in this process were previously adapted to antibiotic effects. To confirm this, the cumulative production of CH 4 was enhanced. This means that antibiotic-contaminated manure non used for AD, and left exposed to air, can produce more methane in the anaerobic inner part of the piles, increasing climate-altering gas emissions. Moreover, because SMX and CIP decreased substantially in digestate, the latter is more desirable as an organic fertilizer in order to reduce the agricultural soils antibiotic contamination and presumably antibiotic resistant genes.
Finally, the antibiotics influenced the structure of the microbial community since Bacteria increased and Archaea decreased with the rise in antibiotic concentrations. In particular, at the end of the experiment, a predominance of archaeal genera (Methanobrevibacter and Methanosphera), effective in the production of methane along the hydrogenotrophic metabolic pathway, was observed. In contrast, the presence of antibiotics counteracted the functionality of acetotrophic methanogens, as confirmed by the acetic acid increased concentrations detected until the end of the experiment.