Characterization of Clostridium Perfringens Isolates Collected from Three Agricultural Biogas Plants over a One-Year Period

Digestate produced by agricultural biogas plants (BGPs) may contain pathogenic bacteria. Among them, Clostridium perfringens deserves particular attention due to its ability to grow under anaerobic conditions and persist in amended soil. The aim of this study was to examine the potential pathogenicity and the antimicrobial resistance of C. perfringens in manure and digestate collected from three agricultural biogas plants (BGPs). A total of 157 isolates (92 from manure, 65 from digestate) were screened for genes encoding seven toxins (cpa, cpb, etx, iap cpe, netB, and cpb2). The 138 cpa positive isolates were then screened for tetA(P), tetB(P), tet(M), and erm(Q) genes and tested for antimicrobial susceptibility. The toxinotypes identified in both manure and digestate were type A (78.3% of the isolates), type G (16.7%), type C (3.6%), and type D (1.4%), whereas none of the isolates were type F. Moreover, half of the isolates carried the cpb2 gene. The overall prevalence of tetA(P) gene alone, tetA(P)-tetB(P) genes, and erm(Q) gene was 31.9, 34.8, and 6.5%, respectively. None of the isolates harbored the tet(M) gene. Multiple antimicrobial resistant isolates were found in samples that were collected from all the manure and digestates. Among them, 12.3% were highly resistant to some of the antibiotics tested, especially to clindamycin (MIC ≥ 16 µg/mL) and tilmicosin (MIC > 64 µg/mL). Some isolates were highly resistant to antibiotics used in human medicine, including vancomycin (MIC > 8 µg/mL) and imipenem (MIC > 64 µg/mL). These results suggest that digestate may be a carrier of the virulent and multidrug resistant C. perfringens.


Introduction
The number of biogas plants has increased considerably in Europe in the last decade (from 6227 installations in 2009 to 18,202 in 2018) [1], most of them operating with agricultural substrates. On-farm anaerobic digestion converts livestock manure into biogas and digestate, and the latter is commonly applied to agricultural soils. However, the spread of zoonotic pathogens and antibiotic resistant bacteria that are potentially present in digestate raises serious concerns. Among these bacteria, Clostridium perfringens, which causes food poisoning in humans and enteric diseases in domestic animals [2], deserves particular attention due to its persistence in soil amended with manure, digestate, or manure compost [3][4][5]. Clostridia are also suspected of playing a role in the dissemination of antibiotic resistance genes in manure-amended soil [4][5][6]. Moreover, C. perfringens has been suggested to be a reservoir for conjugative antimicrobial resistance genes [7,8].
The land application of manure may result in continuous exposure of the agricultural environment to pathogenic and antibiotic resistant bacteria [5]. This raises issues of human health and food safety [9]. Indeed, once contaminated livestock effluents are applied, direct transfer of pathogens to crops and animals may occur [10]. To date, there is no evidence of environmental transmission of C. perfringens, whether or not the bacterium is resistant to antibiotics but one possible transmission route to humans or animals is through the consumption of vegetables contaminated by livestock manure [11,12]. Whether anaerobic digestion of manure can reduce the environmental exposure to antimicrobial-resistant C. perfringens shed by farm animals has to be dully addressed. Given the increasing number of reports of antimicrobial resistance in anaerobes over the past two decades [13], it is important to examine the potential pathogenicity and antimicrobial profiles of C. perfringens present in digested manure before its application on agricultural land.
C. perfringens is a Gram-positive anaerobic spore forming bacterium that is responsible for various diseases, including food poisoning, gas gangrene, enteritis necroticans, and enterotoxemia [14]. It is also involved in antibiotic-associated diarrhea [15]. This bacterium is estimated to be the third most common bacterial cause of food-poisoning cases in the United States [16], and is the third most common cause of foodborne outbreaks in France [17]. C. perfringens produces more than 20 toxins and hydrolytic enzymes [18,19]. The species is divided into seven types, A to G, according to the presence and combination of four major toxins (α, β, ε and ι) encoded by genes cpa, cpb, etx, and iap, and two minor toxins (enterotoxin CPE and NetB) encoded by genes cpe and netB [15,19]. It also produces the β2 toxin encoded by the cpb2 gene, which is implicated in several animal gastrointestinal diseases [20] and it is found in different C. perfringens toxinotypes [18].
Besides carrying toxin genes, the plasmids harbored by C. perfringens also carry genes conferring resistance to antibiotics. The resistance of C. perfringens isolates to tetracycline, clindamycin, lincomycin, and erythromycin has been reported in animals [21,22]. Tetracycline resistance plasmids are the most common antibiotic resistance plasmids in C. perfringens [23]. The genes tetA(P) and tetB(P) are frequently associated with tetracycline resistance [24,25] and the genes erm(B) or erm(Q) have been associated with macrolide-lincosamide-streptogramin B resistance [23,26], the erm(Q) gene being the most common erythromycin resistance determinant reported in C. perfringens. Only a few studies have examined the antimicrobial susceptibility of C. perfringens isolates that were collected from agricultural environments. High resistance to tetracycline has been reported in isolates from piglets, chickens, or cattle [24,[27][28][29]. Moreover, 63% and 38.4% of C. perfringens isolates collected from soil in the United States [28] and in Greece [30] were resistant to tetracycline. Soge et al. [8] also reported that the tetA(P) gene was common in isolates in water and soil. Despite the ability of C. perfringens to transfer antibiotic resistance via bacterial conjugation [7] and its ability to survive under conditions found in mesophilic anaerobic digesters (MAD) [31,32], information is lacking on the antimicrobial susceptibility of resistant C. perfringens isolates in digestate intended for land application.
The aims of this study were (i) to identify the toxinotype and the antimicrobial resistance profiles of C. perfringens isolates in digestates originating from three agricultural biogas plants and (ii) to compare the profiles with C. perfringens isolates in manure that was collected in the same biogas plants.

Biogas Plants and Sample Collection
The samples were collected from three on-farm biogas plants (BGPs) located in a same area (within a distance of less than 120 km), fed with manure and co-substrates of vegetable origin, the proportion of which depending on the BGP (Table S1). BGP1 and BGP3 were fed with liquid pig manure. BGP2 manure mainly comprised liquid dairy manure (≥86%) and to a lesser extent, solid poultry manure.
Each BGP was sampled eight times (at six-week intervals) from May 2017 to July 2018. Prior to sampling, the liquid manure in the storage tank and in the digester was mixed. At each sampling date, liquid manure and raw digestates were collected using three 10 L buckets for each type of matrix.
Samples from each bucket were transferred into 1 L sterile flasks. A total of 48 samples per BPG were analyzed (2 matrices × 3 replicates × 8 sampling dates). Moreover, as poultry manure was not systematically present in BGP2, only three samples were collected during the sampling period. Microbiological analyses were performed within 4 h after sampling.

Isolation of C. perfringens
Samples (25 g) of manure or digestate were homogenized in 225 mL of peptone water. Tenfold serial dilution of the suspensions was performed. C. perfringens was isolated according to the standard ISO method 7937:2005-02 [33]. Briefly, Trytone sulfite Cycloresrin agar (TSC, Thermo Fisher Scientific, Courtaboeuf, France) was poured into a Petri dish containing 1 mL of each serial dilution. After the solidification of the agar, an overlayer of 10 mL of TSC was added. The plates were incubated under anaerobic conditions while using anaerobic jars with gaspack (AnaeroGen, Thermo Scientific, Courtaboeuf, France) at 37 • C for 20 h ± 2 h. Five black colonies were randomly selected on a plate containing between five and 10 colonies of presumptive C. perfringens and transferred into thioglycolate broth (Thermo Fisher Scientific, Courtaboeuf, France). The tubes were incubated under anaerobic conditions at 37 • C for 18 h ± 2 h. Five drops of each thioglycolate culture were transferred into Lactose Sulfite Broth (Biokar Diagnostics, Allonne, France). After incubation at 46 • C for 18 h ± 2 h in a water bath, the tubes exhibiting gas production and blackening of the culture medium were considered positive. At each sampling date, all of the confirmed colonies were purified and suspended in 50% glycerol/50% Brain Heart Infusion Broth (BHI, Thermo Fisher Scientific, Courtaboeuf, France) before being stored at −80 • C. A total of 157 isolates were collected from the three BGPs (Table 1).

DNA Extraction
All of the isolates were cultured on Wilkins-Chalgren agar (WICH, Thermo Fisher Scientific, Courtaboeuf, France). After incubation at 37 • C for 24 h under anaerobic conditions, ca. 40 mg of pure culture were removed with a sterile swab and transferred into a 2 mL tube. DNA extractions were performed using the Nucleospin ® Microbial DNA kit (Macherey-Nagel, Hoerdt, France) according to the supplier's recommendations. The DNA was eluted from the silica column in a 100 µL elution buffer (5 mM Tris/HCl, pH 8.5) and then stored at −20 • C. DNA purity and concentration were determined by measuring absorption at 260 and 280 nm while using a BioPhotometer Spectrophotometer (Eppendorf, Montesson, France).

Determination of the Toxin Genotype and Search for Antibiotic Resistance Genes
Real-time PCR targeting the four major toxin genes (cpa, cpb, etx, and iap), and three minor toxin genes (cpe, netB, and cpb2) was carried out to genotype the C. perfringens isolates ( Table 2). Strain 103,009,089 from the ANSES (French Agency for Food, Environmental and Occupational Health and Safety) strain collection and strains CIP 103409 T , CIP 106527, CIP 105568, CIP 106156, and CIP 104612 from the Institut Pasteur Collection were used as positive controls.
Real-time PCRs were performed using a Bio-Rad CFX96 real-time PCR machine with Bio-Rad CFX Manager software, version 1.1 (Bio-Rad, Marnes-la-coquette, France) in a final volume of 25 µL. The reaction mixture consisted of 12.5 µL of IQ SYBR Green Supermix (Bio-Rad, Marnes-la-coquette, France), a 400 nmol/L concentration of each primer, 2 µL of 1/10 diluted DNA, and 9.5 µL of water. Negative controls without DNA were also prepared. PCR preparation was performed while using the epMotion 5070 automated pipetting station (Eppendorf, Montesson, France).
The Real-time PCR program for the target gene included one cycle at 94 • C for 5 min, 40 cycles at 94 • C for 20 s, one cycle at 55 • C for 30 s, and one cycle at 72 • C for 30 s [34,35]. The melting curve of the amplification products was analyzed at the end of each qPCR (temperature range 65 • C to 95 • C). Analysis of the melting curve of the standard replicates only yielded one peak resulting from the specific amplification of the target gene. For a given gene, the amplification of the correct PCR fragment was verified by comparing the cycling threshold value and the melting curve of the standard PCR product with each environmental isolate PCR product.

Testing Antimicrobial Susceptibility
Antimicrobial susceptibility was tested on the 138 cpa-positive isolates using the commercially available Sensititre™ Bovine/porcine according to the manufacturers' instructions (ThermoFisher Diagnostics, Courtaboeuf, France). Seventeen antimicrobial agents commonly used for the treatment of livestock diseases were tested. They included (µg/mL) penicillin (PEN, 0. The isolates were incubated in WICH broth at 37 • C in under anaerobic conditions. Overnight bacterial cultures were adjusted to 0.5 McFarland turbidity standard in WICH broth, and 50 µL of the diluted bacterial suspensions (ca. 10 5 CFU/well) were added to each well.
Among the 138 cpa-positive isolates, 14 isolates originated from the three BGPs were selected for their high resistance to antibiotics of the Sensititre™ Bovine/porcine plate. Their susceptibility to antimicrobial agents commonly used against anaerobic human pathogens were tested with the Sensititre™ Anaerobe MIC plate (WICH, Thermo Fisher Scientific, Courtaboeuf, France) according to the manufacturers' instructions (ThermoFisher Diagnostics, Dardilly, France). The plate  Tables S2 and S3.
The lowest concentration at which no bacterial growth was detected corresponded to the minimum inhibitory concentration (MIC) value. MIC 50 and MIC 90 were the MIC values at which 50% and 90% of the isolates were inhibited.

Statistical Analysis
The MIC of the isolates for the tested antimicrobials was subjected to principal components analysis (PCA) to explore the relationship between the origin of the isolate (BGP, manure, digestate) and their antimicrobial resistance profiles. All of the statistical tests were performed with XLSTAT 2019 (Addinsoft, Bordeaux, France). Heatmap of normalized antimicrobial resistance were generated using the Complex Heat map package in Circlize in R [36].

Toxinotyping of C. perfringens Isolates
Among the 157 isolates, 138 (88%) carried the cpa gene (43 from BGP1, 47 from BGP2, and 48 from BGP3). As shown in Table 3, 108 (78.3%) were type A, and 23 (16.7%) were type G (netB gene). Both types were found in pig and dairy manure and in the digestate in all three BGPs. Five isolates (3.6%) and two isolates (1.4%), found only in BGP3, were type C (cpb gene) and type D (etx gene), respectively. The proportion of C. perfringens isolates harboring the cpb2 gene were 48.1%, 50%, 56.5%, and 80% for types A, D, G, and C, respectively. None of the isolates harbored both cpe and iap genes.

Tetracycline and Erythromycin Resistance Genes of C. perfringens Isolates
Both tetA(P) and tetB(P) genes were detected in the 138 cpa-positives isolates from manure and digestate of the three BGPs, but not in isolates from poultry manure (Table 4), whereas the erm(Q) gene was only found in isolates from BGP1 and BGP3. Table 4. Detection of antimicrobial resistance genes (tetA(P), tetB(P), tet(M), and erm(Q) genes) of isolates in manure and digestate in three biogas plants.  The most frequent gene, tetA(P), was found in 62.5% and 86.2% of the isolates from the manure and digestate in the three BGPs, respectively. The tetB(P) gene was associated with tetA(P) genes in 31.3% and 50% of the isolates from manure and digestate, respectively. The erm(Q) gene was found less frequently, i.e., in 3.8% of the isolates from manure and 10.3% of the isolates from digestate. No tet(M) gene was detected in the isolates.
The antimicrobial resistance profiles of the isolates varied considerably in the biogas plants, regardless of their origin (manure or digestate) ( Figure S1). However, the MIC distribution of isolates from manure was close to that of isolates from digestate (Table 5, Figure S2).
With the exception of tulathromycin, the MICs of the isolates covered the dilution range tested for all antimicrobials ( Table 5). Most of the isolates were susceptible to β-lactam and macrolides, except tulathromycin (with a MIC 50 value of 64 µg/mL). Enrofloxacin and danofloxacin had low MIC 50 values of 0.25-0.5 and 1 µg/mL, respectively. Most of the isolates exhibited sensitivity to chlortetracycline with a MIC 50 value of ≤0.50 µg/mL in manure and of 1 µg/mL in digestate. The MIC 50 values of oxytetracycline were higher (1 µg/mL in manure and 8 µg/mL in digestate).
Moreover, the growth of 19% of C. perfringens isolates was inhibited by sulfadimethoxine ( Figure S1). Among the highly resistant group I to V, 14 isolates were tested on microplates that contained antibiotics used in human medicine (Table S3). Five of the tested isolates present in BGP2 and BGP3, showed high resistance to human antibiotics including imipenem, linezolid, amoxicillin-clavulanate, chloramphenicol, moxifloxacin, tigecycline, and vancomycin. The antimicrobial resistance profiles of the isolates varied considerably in the biogas plants, regardless of their origin (manure or digestate) ( Figure S1). However, the MIC distribution of isolates from manure was close to that of isolates from digestate (Table 5, Figure S2). With the exception of tulathromycin, the MICs of the isolates covered the dilution range tested for all antimicrobials ( Table 5). Most of the isolates were susceptible to β-lactam and macrolides, except tulathromycin (with a MIC 50 value of 64 µ g/mL). Enrofloxacin and danofloxacin had low MIC 50 values of 0.25-0.5 and 1 µ g/mL, respectively. Most of the isolates exhibited sensitivity to chlortetracycline with a MIC 50 value of ≤0.50 µ g/mL in manure and of 1 µ g/mL in digestate. The MIC 50 values of oxytetracycline were higher (1 µ g/mL in manure and 8 µ g/mL in digestate).
Moreover, the growth of 19% of C. perfringens isolates was inhibited by sulfadimethoxine ( Figure  S1). Among the highly resistant group I to V, 14 isolates were tested on microplates that contained antibiotics used in human medicine (Table S3). Five of the tested isolates present in BGP2 and BGP3, showed high resistance to human antibiotics including imipenem, linezolid, amoxicillin-clavulanate, chloramphenicol, moxifloxacin, tigecycline, and vancomycin.

Discussion
Although digestates from biogas plants fed with manure contain high levels of C. perfringens (ca. 10 4 CFU/g) [32][33][34][35][36][37], no information is available on the range of toxin genes and the antimicrobial profiles of this pathogenic bacterium in agricultural digestate that is intended for land application. This is the first study to assess digestate as a potential carrier of virulent and multiresistant C. perfringens.

Toxinotyping of C. perfringens Isolates
The dominance of type A isolates in both manure and digestate is in agreement with the high prevalence of this toxinotype in farm animals [38][39][40]. Although the cpe gene was not detected in either manure or digestate, the three BGPs produced digestate containing virulent C. perfringens. Indeed, half the type A isolates from the three BGPs carried the cpb2 gene encoding the β2 toxin, which is reported to be associated with diarrheal diseases [19,41]. This result is consistent with those of studies that were conducted on farm animals, in which between 39% and 78.8% frequency of the cpa gene together with cpb2 gene was observed [25,[42][43][44]. Moreover, some isolates recovered from manure and digestate were positive for cpb, etx, and netB genes. Conversely, Chan et al. [45] did not detect these genes in fecal and manure samples collected from pig farms. Type C (cpa + cpb genes) and type D (cpa + etx genes), both associated with farm animal diseases, were detected at a low frequencies only in isolates from the BGP3 digestate. Type G (cpa and netB genes), responsible for necrotic enteritis in chickens [6,29,45], was carried by isolates from all three BGPs, showing that land application of digestate can contribute to the dissemination of C. perfringens carrying netB in the agricultural environment.

Tetracycline and Erythromycin Resistance Genes of C. perfringens Isolates
The occurrence of resistance genes and the antimicrobial susceptibility of C. perfringens of environmental origin have rarely been studied. In this study, although the tet(M) gene was not detected in digestate, the tetA(P) gene alone and double resistance tetA(P)-tetB(P) genes were found in 34.5% and 41.4% of the isolates, respectively. Their prevalence in digestates was slightly higher than that reported in other matrices. Indeed, Kiu et al. [14] observed a frequency of detection of tetA(P) alone and of the tetA(P)-tetB(P) gene of 14.8% and 35.2% in isolates from broiler chickens, respectively. In another study, they were detected in 26% and 22% of isolates of environmental origin (soils, water, and sewage) [8]. The proportion of isolates carrying the erm(Q) gene (13.6%) in the digestates of the two BGPs fed with pig manure was also higher than that previously observed in isolates that were collected from broiler chickens (4.5%) and in environmental samples (1%) [8,14].

Antimicrobial Susceptibility of C. perfringens Isolates
In addition to their natural resistance to aminoglycosides, C. perfringens showed significant resistance to tetracycline, which has been associated with the presence of tetA(P), tetB(P), and tet(M) genes [27]. In this study, the isolates were more resistant to oxytetracycline than to chlortetracycline (Table 5), in agreement with the studies of Martel et al. [46] and Ngamwongsatit et al. [38] who reported that chlortetracycline was more active than oxytetracycline against C. perfringens isolated from chickens and from neonatal piglets, respectively. It is noteworthy that 7.4% of the isolates with a MIC ≥ 4 µg/mL for chlortetracycline and with a MIC ≥ 8 µg/mL for oxytetracycline harbored none of the three target genes (tetA(P), tetB(P), and tet(M)). This could be explained by the presence of other tetracycline resistance genes. Indeed, tet(K) and tet(L) have been detected at low frequencies in isolates from broilers [46,47]. Moreover, 4.4% of the isolates that were susceptible to both chlortetracycline and oxytetracycline (MIC ≤ 0.5 µg/mL) carried the tetA(P)-tetB(P) genes and 22.2% carried the tetA(P) gene. The inconsistency between the MIC values and the presence of genetic tetracycline resistance determinants has already been pointed out [7,24,47]. The reason for the presence of sensitive isolates harboring tetA(P) and/or tetB(P) is unclear. This may be due to lack of expression or to point mutation, as suggested by Kather et al. [7] and Gholamiandehkordi et al. [47]. In the present study, all strains were highly resistant to tulathromycin, a macrolide antibiotic used to treat bovine and swine respiratory diseases [48]. However, it was not possible to determine the antibiotic use in the three farms whose manure was treated by anaerobic digestion. On the whole, the MIC 50 observed in manure and digestates for the antibiotic tested resembled those reported in animal feces [22,27,29,38,49,50], suggesting that storing manure and anaerobic digestion have no impact on the antimicrobial susceptibility of C. perfringens originating from farm animals.
Among the C. perfringens isolates from the three BGPs, 12.3% were highly resistant to some of the antibiotics tested. Five of the 14 isolates tested with the Sensititre™ Anaerobe MIC plate (four from BGP3, one from BGP2) were also highly resistant to human antimicrobial medicine (Table S3). Based on the values of the EUCAST anaerobe MIC breakpoints reported by Pence et al. [13], these isolates were resistant to metronidazole, clindamycin, and penicillin (MIC ≥ 8 µg/mL), to chloramphenicol, amoxicillin-clavulanate, and imipenem (MIC ≥ 16 µg/mL), and to piperacillin-tazobactam (MIC > 64 µg/mL). High MIC were also observed for vancomycin (MIC > 8 µg/mL), which is normally low for clostridia, and for moxifloxacin (MIC > 4 µg/mL), the resistance to which is reported to be increasing [13].
In this study, digestates were collected from digesters fed with manure, which is the first step of the process of anaerobic digestion. Indeed, post-treatments of raw digestates involved either storage in a tank, post-digestion (BGP2), or mechanical separation of the solid fraction before composting and storage of the liquid fraction (BGP1). It has been reported that post-digestion and mechanical separation do not reduce the concentration of C. perfringens [37], contrary to manure composting [3,51]. Moreover, the legal requirement for heat treatment (70 • C for 1 h) according to the EU animal by-products regulation (EU Commission Regulation (EU) No 142/2011) is not sufficient for eliminating spores of C. perfringens. Indeed, Sahlström et al. [52] reported that the number of C. perfringens was not significantly affected by this time-temperature condition. On the opposite, high temperatures (135 • C to 190 • C) pre-treatments performed in order to increase accessibility of the enzymes to the substrate and, therefore, to enhance biogas production [53,54], may reduce the concentration of C. perfringens. Actually, the inactivation of C. perfringens spores can be achieved at temperatures of 110 • C [55]. Thus, high temperature pre-treatment or digestate composting may be applied to avoid C. perfringens spreading in soils.

Conclusions
The results of this study revealed that both manure and digestates contained C. perfringens isolates belonging types C, D, and G, which may be pathogenic for human or animals. Moreover, the presence in the digestates of some isolates highly resistant to antibiotics used in both veterinary and human medicine may increase the burden of antimicrobial resistance in agroecosystems, as C. perfringens could be involved in horizontal gene transfer from enteric bacteria to bacterial soil populations.
Supplementary Materials: The following are available online at http://www.mdpi.com/1660-4601/17/15/5450/s1, Figure S1: Heatmap of antimicrobial resistance profiles of C. perfringens isolates from three biogas plants, Figure S2: MIC distribution of 13 antimicrobials in manure and raw digestate in three BGPs, Table S1: Process parameters of the three biogas plants, Table S2: MIC of C. perfringens CIP 103409 T obtained on Sensititre TM Bovine/porcine (µg/mL), Table S3: MIC of selection of isolates (belonging to groups I to V) from the three BGP and of the reference strain CIP 103409 T obtained on Sentititre TM FRAM1ANA (µg/mL). Funding: This research was funded by the French Environment and Energy Management Agency (ADEME), grant number 1606C0022.