Genotyping of Clostridium perfringens Isolates from Domestic Livestock in Saudi Arabia

The present study was undertaken to confirm the genetic identity of Clostridium perfringens isolates from domestic livestock in Saudi Arabia and to characterize the genes encoding to alpha, beta, epsilon, and iota (α-, β-, ε-, and ι-) toxins. C. perfringens were confirmed in 104 out of 136 isolates on multiplex polymerase chain reaction using specific primers amplifying genes related to toxins produced by C. perfringens. Genes encoding α-toxins were detected in 104 samples. Of the isolates, 80.8% were diagnosed as type A, 15.4% as type D, 2.9% as type C, and 0.96% as type B. None of the isolates has genes encoding iota (ι-) toxin. All isolates investigated yielded enterotoxin (cpe) products and none yielded β2 (cpb2-toxin) or NetB products. PLC gene sequences encoding α-toxin showed >96.7% similarity. Isolates which had α-toxins as well as enterotoxin (cpe) are regarded as type F. Phylogenetic analysis using maximum likelihood analysis yielded two clades, and the majority of the isolates were in one group while only two isolates clustered on the second clade. Within the Kingdom of Saudi Arabia strains, 54 variable positions and 23 polymorphic amino acids were noticed. Isolates with ε- and β-toxins were variable and were found to be close to those published for C. perfringens. ETX gene sequences encoding ε-toxins were found to be related to CPE sequences.


Introduction
The anaerobe and spore-forming Clostridium perfringens is a Gram-positive rod-shaped bacterium detected in soil and water as well as among intestinal bacterial flora of animals and man [1]. It is categorized with the most widespread bacteria, which occurs in different environments such as soil, sewage, food, feces, and the normal intestinal flora of humans and animals [2]. The toxins produced by different species of Clostridium are more than those produced by other eubacteria [3]. All strains of C. perfringens possess the gene encoding for phospholipase C (plc) (also referred to as alpha toxin α) in combination with differential expression of 3 major toxin-encoding genes (beta β, epsilon ε, and iota ι) used to classify strains as toxinotypes A to E [4]. C. perfringens isolates from type A produce only α-toxin, those from type B produce α-, β-, and ε-toxins, those from type C secrete αand β-toxins, those from type D produce αand εtoxins, and bacteria from type E produce αand ι-toxins [5]. C. perfringens is a human and livestock pathogen, resulting in disease in different tissues as well as intestinal diseases in the form of enterotoxaemia. C. perfringens type A toxigenic group causes food poisoning in humans and enteritis in domestic and wildlife animal species, type B and C cause necrotizing enteritis and enterotoxaemia in cattle, sheep, and other animals, type D affects sheep, goats, and cattle mainly, and type E causes enteritis in rabbits, sheep, and cattle [6,7]. Recent results by Rood and others suggested a modification for the toxinotyping of C. perfringens which required the inclusion of the enterotoxin (CPE) and NetB toxins produced by the bacterium [8]. They proposed that strains which produce α-toxin as well as CPE are assigned type F whereas those producing α-toxin and NetB toxins are assigned type G. In both cases, no other toxins such as β-, ε-, and ι-toxins are produced.
The prevalence of C. perfringens in Saudi Arabia was studied by some investigators [9,10]. In 2006, during the annual Hajj pilgrimage, a group of Saudi soldiers were diagnosed for having gastroenteritis following a rice lunch contaminated with Bacillus cereus and C. perfringens [9]. Moussa and Hessan [10] detected C. perfringens in 20% of the calves, sheep, and poultry investigated. C. perfringens was detected in the dromedary camel in the eastern region of Saudi Arabia, where a prevalence of 23.4% was reported from the samples investigated [11].
Diagnosis of C. perfringens enterotoxaemia in various animal species is dependent on clinical signs as well as pathological findings; however, characterization of toxins in intestinal contents is of importance to confirm the clinical diagnosis.
The widely used identification of the toxins is based on mice neutralization test, but this method is expensive, timeconsuming, and subjecting laboratory animals to stress and unnecessary risk of being affected by disease [12]. Enzymelinked immunosorbent assay (ELISA) kits have also been used for the detection of clostridial toxins [13,14]. DNAbased techniques, such as the polymerase chain reaction (PCR) was developed for C. perfringens genotyping which is a reliable alternative method to replace testing in laboratory animals. Different PCR protocols were employed to genotype C. perfringens isolates with respect to the genes cpa, cpb, etx, iap, cpe, NetB, and cpb2, which encode the α-, β-, ε-, ι-toxins, CPE, NetB, and β2-toxins, respectively [4,15].
In the present study, C. perfringens isolated from livestock in Saudi Arabia has been characterized and the sequence variation in α-, ε-, and β-toxins producing isolates has been investigated.

Materials and Methods
Isolates of C. perfringens were obtained from field investigations for the diagnosis of enterotoxaemia in domestic livestock in Saudi Arabia. Intestinal swabs and intestinal contents were taken to the laboratory of the Ministry of Environment, Water and Agriculture (MEWA), KSA. In the laboratory, samples were subjected to microscopic examination using Gram stain as well as to serological testing. Representative isolates were subjected to DNA extraction, polymerase chain reaction, and DNA sequencing.

Microscopic Examination of Gram-Stained Smears.
Impression smears were prepared from parts of the small intestines and smears were also prepared from suspected bacterial isolates. Smears were air-dried, heat fixed, stained with Gram stain, and examined microscopically for detection of large-sized Gram-positive bacilli similar to clostridia.

Detection and
Typing of C. perfringens by ELISA Assay. The Bio-X enterotoxaemia ELISA kit (Bio-X Diagnostics, Belgium) was used to detect α-, β-, and ε-toxins, cellular antigens of C. perfringens in clinical samples, and bacterial cultures according to the manufacturer's instructions. Typing of C. perfringens isolates into the different types A, B, C, or D was done by comparing the types of toxins detected in the clinical material or culture supernatants of pure colonies of C. perfringens isolates grown in the liquid Trypticase Glucose Yeast (TGY) extract medium. The TGY medium is composed of Tryptone, yeast extract, glucose, K 2 HPO 4 , agar, and water.
2.3. Bacterial DNA Extraction, Polymerase Chain Reaction (PCR), DNA Sequencing, and Phylogenetic Analyses. DNA was extracted from 136 supposedly C. perfringens isolates based on the microscopic examination, using the QuickEx-tract™ Bacterial DNA Kit from Epicentre® (an Illumina® Company). Polymerase chain reaction (PCR) and multiplex PCR were performed using the primers illustrated in Table 1. The plc gene was PCR amplified from extracted DNA or cDNA using the primer pair designed by Baums and others [15]. Amplifying NetB gene was performed using primers APK78-F and APK79-R as indicated by Keyburn and others [16].
The plc (cpa), cpb, ext, ipa, cpe, NetB, and cpb2 genes were amplified in 25 μl volumes containing 5 μl PCR buffer, 1 pmol each specific primer, 200 mM each dNTP, 4 U MyTaq DNA polymerase (Bioline, London, UK), and 2 μl of DNA extract. Thermal cycling conditions were as follows: 2 min at 94°C, followed by 35 cycles of 94°C for 30 sec, 55°C (up to 58°C depending on the primer) for 30 sec, 72°C for 30 sec, and a final extension at 72°C for 5 min. The PCR products were visualized using stained ethidium bromide 1.5% agarose gel exposed to ultraviolet light, and digital images were taken from the PCR product through gel documentation system. Positive PCR products obtained were purified using the QIAquick kit (QIAGEN, Hilden, Germany) according to the manufacturer's instructions.
Purified PCR products were sequenced using the Macrogen facility, Seoul, South Korea. The plc gene sequences were manually aligned according to their nucleotides and the deduced amino acid sequences using MEGA X, and all alignments were refined by manual inspection.
Phylogenetic analyses were conducted using the computer program MEGA X [17]. The best-fit substitution model was TN93+G+I. The statistical reliability of internal branches was assessed from 1000 bootstrap pseudoreplicates. Phylogenetic trees were generated in MEGA X using the maximum likelihood (ML) approach.
The deduced amino acids from the coding regions were obtained using MEGA X. Then, they were compared with the existing sequences in the NCBI Protein database to identify homologous sequences through PSI-BLAST. Sequences from different strains were aligned with the ClustalW program using the MAFFT, a multiple sequence alignment program [18]. The secondary structure of plc (α-helix and β-sheet) was predicted using the PDBSAS server [19]. Relevant annotations were mapped using a consensus sequence recovered from the PDB server. Phylogenetic trees were generated using the maximum likelihood (ML) analysis in MEGA X.

Results
Clinical samples collected for bacteriological identification yielded C. perfringens based on cultural, microscopic, and cellular antigenic characterizations. Molecular characterization of 136 isolates showed that 104 isolates yielded positive PCR products. Of those 138 isolates, 84 (80.8%) had only the plc gene (type A), 16 (15.4%) isolates had the etx gene as well as the plc gene (type D), 3 (2.9%) isolates had the cpb gene as well as the plc gene (type C), and 1 (0.96%) isolate had the plc, cpb, and etx genes (type B). All isolates investigated in the present study had the cpe gene; however, none had iap, cpb2, or NetB genes. Figure 1 shows representative isolates which showed positive PCR products in a multiplex PCR. The expected sizes of the PCR product from different primers are shown from representative samples.
3.1. Alpha (α-plc) Gene Characterization. A total of 48 sequences from alpha toxins were analyzed. Sequences of alpha toxins obtained in the present study were highly similar, and they showed a similarity of >96.7%. There was no specific sequence type associated with a known location. The sequences investigated were from 5 different cities including Riyadh, Hofuf, Jeddah, Taif, and Qassim. The dataset used in the analysis included 104 sequences. Of those, 48 were new sequences (present study), 22 were published from animals and humans in the USA [20], 32 strains were isolated from Danish broiler chickens grouped according to pulsedfield gel electrophoresis (PFGE) profiles [21], one unpublished sequence (MK599266) was from camel, and one was out-group Clostridium novyi (D32125).
The ML topology showed two major clades with high bootstrap values, the first is divided into 10 subclades including 46 of the isolates. The most divergent isolates, 1 and 7, were localized in the second clade ( Figure 2). The strain MK599266 grouped with strains from Hofuf and Riyadh.
The plc gene was reported to be highly conserved in C. perfringens strains. We observed 275 and 47 variable positions, respectively, out of the 860 nucleotides and 272 deduced amino acids. The studied protein-coding region displayed moderate-level nucleotide diversity (Pi = 0:019). The dataset generated 74 haplotypes characterized by a high haplotype diversity (Hd = 0:987). Within KSA samples, we observed 54 variable positions and 23 polymorphic amino acids ( Figure 3).
α-Gene sequences obtained in the present study were compared to similar sequences obtained from the GenBank. The sequences from the Kingdom of Saudi Arabia were scattered and clustered with sequences obtained from Denmark and USA ( Figure 2). Fourteen sequences were clustered in Table 1: Primers used for amplification of the genes responsible of producing specific toxins and the expected amplicon size together with the annealing temperature for the PCR.

Gene
Primer sequence 5′ to 3′ Amplicon size (bp) Annealing temp (°C)    BioMed Research International one group with some other sequences from the GenBank while other sequences were scattered between other sequences. Protein sequences differed in 22 positions between isolates from Saudi Arabia, and they showed 31 different types. They showed differences in 41 amino acid position when compared with isolates from USA and Denmark; however, some of them have shown similarities (Figure 3). The sequence variation was not related to the pathogenicity of the strain as all organisms investigated in this study were obtained from pathological cases. The secondary structure prediction revealed that plc amino acid sequences of α-toxin mainly consists of alphahelices ( Figure 4). The α-toxin enzyme contains an α-helical N-terminal domain (residues 1-246) encompassing the active site and a β-sandwich C-terminal domain (residues 247-255). The β-sandwich domain includes a calciumbinding pocket (Figure 4). The Protein Data Bank (PDB) multiple alignment confirmed the structural composition and showed 9 proteins with identity percentage varying between 95.4% and 98.9%, all α-toxin belonging to C. perfringens. Clostridium absonum α-toxin showed only 62% identity compared to those obtained from C. perfringens.

Beta (β-) Gene
Characterization. The dataset included 16 sequences comprising 4 new and 12 reference sequences published previously (HQ111476-82 and KU836730 from Iran, L13198 from the United Kingdom, and X83275 from Iceland). Annotation of the obtained sequence revealed that the region comprising between 63 and 68 corresponds to a regulatory class="ribosome binding site." The fragment from the position 76 to 560 corresponds to a partial product of the β-toxin. Multiple alignments of cpb gene sequences yielded 21 variable sites, and only three were in the coding region.
3.4. Enterotoxin (cpe) Production. All isolates showed positive results for the C. perfringens enterotoxin (cpe). Different haplotypes were reported; however, the majority of the haplotypes were identical to those reported for C. perfringens available in the GenBank.

Discussion
The identity of Clostridium perfringens isolated from domestic livestock in Saudi Arabia has been confirmed using molecular techniques. Furthermore, different genes responsible of excreting each toxin have been identified and characterized. C. perfringens type A has been reported in 80.8% of the bacterial isolates, type D was reported in 15.4%, type C was isolated from 2.9%, and type B was reported in 0.98%. This finding was similar to a previous work which recorded the four types of C. perfringens reported in this study but with variable rates of prevalence [11]. They reported C. perfringens type A in 33.3%, type B in 5.3%, and types C and D both in 5.3% of the camel investigated. On an earlier study, C. perfringens types A in 81.5%, B in 3.7%, C in 3.7%, and D in 11.1% of sheep, goats, and chicken investigated in Riyadh region were reported [10]. In both studies, the neutralization test in mice were used and they did not include either cpe or NetB gene in their studies. Organisms which were recognized, in the present study, as type A may be designated as type F rather than type A as all produced CPE toxin together with α-toxin according to the suggestion made by Rood and others [8]. We were able to demonstrate the presence of C. perfringens type F for the first time in Saudi Arabia. Results of the present study indicate that C. perfringens type A (or type F) is the most prevalent type, and these results were also similar to those found in previous studies [22][23][24][25][26][27]. The gene coding to α-toxin in Saudi Arabia was found to show variability, and it is interesting that some of the strains are related to strains isolated from Japan and USA [20,21]. Furthermore, strains which were reported from Riyadh clustered with those obtained from Taif, Jeddah, and Qassim. Such finding can be attributed to the fact that animal movement in Saudi Arabia is inevitable and Riyadh is a big market for livestock; therefore, animals are transported between different cities, and hence, bacterial infection can easily be found in various places.
Epsilon-coding gene sequences were found to be related to cpe genes produced by type C and D isolates from different animal species in the study conducted by some investigators [28,29]. CPE loci were found to be divergent in C. perfringens isolates. All cpe-positive type D isolates studied previously were found to carry both cpe and etx genes on the same plasmid [28,30]. It is likely the case of having etx grouped with cpe in the current study. Such observations  K   I  I  I  I  I N  K  R  I  Q  S  E  G  T  D  D  K S  A  L Y  Y  W  P  I  S  S  F  FW  H  P  N  N  D  T  D  D  D  D Y  Y  A  N  K  D  P  P PSVG K macromolecules are included in the smaller and larger protein complexes which cause paracellular permeability changes [36,37]. The mechanism of action of enterotoxin is binding to the intestinal epithelial cells forming a series of protein complexes in cell membranes, which ultimately results in altering membrane permeability and cell lysis. β2-Toxin was found to be present among cpe-positive C. perfringens type A isolates from humans with antibioticassociated diarrhea. Of 35 cpb2-harbouring enterotoxigenic C. perfringens type A strains, production of β2-toxin was produced in 34 isolates in vitro [38]. The presence of cpb2 was less common among C. perfringens strains isolated from humans suffering from food poisoning [38].
Genotypic differences exist between cpe type A-positive isolates causing food poisoning and those causing nonfoodborne gastrointestinal disease [39]. It has been demonstrated that most, if not all, C. perfringens type A food poisoning isolates carry their cpe gene on the chromosome while most, if not all, CPE-associated nonfoodborne human gastrointestinal diseases carry a plasmid cpe gene [40,41]. CPE sequences reported in the present study showed high similarity to those described from the GenBank, and they are of plasmid origin. This suggested that the cpe sequences reported from all αtoxin-producing strains are probably of plasmid origin and not of chromosomal origin.
Overall, the mutations that were found in the amino acid sequences of α-toxin did not significantly alter the predicted structure of the proteins. The studied isolates of plc showed high sequence identity (>95%) to PDB Clostridium strain. None of the detected amino acid mutations occurred in the active site or in the partial C-terminal domain ( Figure 2). These regions have been shown to play key roles in proteinmembrane interactions prior to phospholipid cleavage [42].

Conclusion
Clostridium perfringens types A, B, C, and D have been reported from the Kingdom of Saudi Arabia. C. perfringens type A has been reported in 80.8% of the bacterial isolates, type D was reported in 15.4%, type C was isolated from 2.9%, and type B was reported in 0.98%. There was some genetic variation in the conserved plc gene which is responsible of the production of alpha toxins. In the present study, 275 and 47 variable positions, respectively, out of the 860 nucleotides and 272 deduced amino acids, were observed. All isolates studied produced CPEs. When considering vaccine production, field isolates which comprise all the strains must be considered.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.