Prevalence of Clostridium Perfringens Spores in Selected Regions of Saudi Arabia

Clostridium perfringens is an anaerobic, gram-positive, spore-forming, enteric bacterial pathogen that causes a wide range of human and animal diseases owing to its prolific toxin-producing capability [1-7]. C. perfringens can be classified into seven types (A through G), based on the presence of genes encoding six major toxins (\(\alpha\), \(\beta\), \(\epsilon\), \(\tau\), C. perfringens enterotoxin (CPE), and C. perfringens necrotic enteritis \(\beta\)-like toxin (NetB)) [1,8]. CPE, the medically important toxin produced by C. perfringens type F, is the major virulence factor for C. perfringens type F food poisoning (FP) and non-foodborne (NFB) gastrointestinal (GI) diseases [9]. C. perfringens type F FP is the third most commonly reported foodborne disease in the United States. The annual cost of illness is estimated to be more than $300 million [10-12]. Interestingly, in C. perfringens type F isolates, the CPE-encoding gene (cpe) can be located either on the chromosome or on a plasmid. In general, chromosomal cpe isolates are generally linked to FP, whereas plasmid-borne cpe isolates are associated with NFB GI diseases [13-15]. Nevertheless, some studies have found that plasmid-borne cpe isolates can also be causative agents for C. perfringens type F FP [15-18].

Numerous studies have attempted to understand why type F isolates carrying a chromosomal cpe gene are more likely to be associated with C. perfringens FP outbreaks [14,16,18,19]. A survey reported that  1.7% of raw meat, fish, and poultry items sold in retail food stores contain type F isolates carrying chromosomal cpe [20]. Interestingly, this survey noted the absence of type F plasmid-borne cpe isolates in retail foods. These findings suggest that meat, seafood, and poultry, common food vehicles for C. perfringens FP in the United States and Europe can be contaminated with type F chromosomal-cpe isolates by the time of retail purchase. However, these survey results do not preclude the possibility of food contamination in food processing environments. C. perfringens spores are more resistant to a number of lethal factors than their vegetative forms [15,21,22]. Spores, especially spores of FP strain (21), can survive thermal processing and sanitizing treatments employed in the food industry. In addition, they are highly hydrophobic, complicating their removal when they are attached to food contact surfaces [23-26]. A potential source of pathogen transmission to food products is the contamination of food contact surfaces during food processing, catering, and in domestic environments [21,26]. Another possibility is that the food items might become contaminated with type F chromosomal-cpe isolates residing in environmental niches, such as the soil or home kitchen surfaces [27,28]. A study that surveyed different soils and home kitchen surfaces in Pittsburgh, PA, did not detect C. perfringens isolates from home kitchens, while most of the soil samples tested positive for this bacterial isolate. The soil isolates were predominantly type A, although types C, D, E and F were also identified. All cpe-positive soil isolates were genotyped as type F, harboring cpe genes on a plasmid [27].

Although Clostridium spore-mediated disease outbreaks, such as FP and Clostridioides difficile infection (CDI), are common in the USA and Europe [21,29], no such outbreaks have been systematically documented in Middle Eastern countries, including the Kingdom of Saudi Arabia (KSA). The absence of disease monitoring and control systems implies that the outbreaks normally occur in Saudi Arabia, but have not yet been reported. The KSA has no data on the incidence or prevalence of Clostridium-associated diseases, except for a limited study in the Dhahran region [30]. Therefore, the purpose of this study was to determine the prevalence of C. perfringens spores in the Hail and Qassim regions of the KSA by: 1) isolating C. perfringens bacteria from various samples obtained from soil, food, and hospital floors; 2) evaluating the presence of plc (encoding \(\beta\)-toxin) and cpe in newly-isolated C. perfringens strains using polymerase chain reaction (PCR); and 3) examining genome organization by determining the DNA sequences of the representative Saudi isolates. Our results suggest that C. perfringens spores are highly prevalent in the Hail and Qassim regions, KSA.

A. Survey of soil, food, and hospital floors for the presence of C. perfringens spores

We selected three hospitals from both Qassim (hospitals A-C) and Hail (hospitals D-F) regions of KSA for our study. We collected soil samples from the hospital surroundings, swab samples from hospital floors, and food samples, such as various raw meats, including ground beef, and chicken supplied to hospitals. Collectively, 300 soil, swab, and food samples were collected from different places.

B. Isolation of C. perfringens from soil samples

Soil (1.0 g) was collected in a 15-ml sterile plastic tube and then 2.0 ml sterile tryptone glucose yeast (TGY) broth (3% trypticase [Difco, BD Diagnostic Systems, Sparks, MD, USA], 2% glucose [Sigma-Aldrich, USA], 1% yeast extract [Difco, BD Diagnostic Systems, Sparks, MD, USA], and 0.1% L-cysteine [Sigma-Aldrich, USA]) was added and mixed vigorously. An aliquot (0.1 ml) of suspension was cultured onto TSC (tryptose-sulfite-cycloserine, a relatively selective medium for Clostridium isolates) (Millipore, Burlington, MA, USA) agar plates and incubated in anaerobic jars containing GasPak (BD EZ anaerobe container system. Becton, Dickinson and company spark, Maryland USA) at 37 ºC for 24 h. For enrichment,   1.0 ml aliquot of TGY-soil suspension was added to each of two tubes containing 9 ml of sterile TGY. One tube was incubated at 37 ºC overnight ( 18 h) to grow vegetative cells. The other tube was heat-shocked at 75 ºC for 20 min, then anaerobically incubated at 37 ºC overnight ( 18 h), allowing spores to germinate and grow. If there was growth in a TGY tube, an aliquot (0.1 ml) of TGY grown culture was plated onto the TSC plate and anaerobically incubated at 37 ºC for 24 h. Three black colonies of the TSC plates were selected and allowed to grow in TGY at 37 ºC for 18 h. These TGY-grown cultures were then streaked onto sheep/horse blood agar plates and anaerobically incubated at 37 ºC for 24-48 h. The culture that produced a clear double zone of \(\alpha\)-hemolysis, with the inner zone (complete hemolysis) caused by perfringolysin O and the outer zone (partial hemolysis) caused by \(\beta\)-toxin, was considered as C. perfringens. C. perfringens cultures were stored as glycerol stock at -80 ºC freezer until used.

C. Isolation from raw meat samples

Twenty-five grams of each raw meat or meat product was suspended in 225 ml 0.1% peptone (Difco, BD Diagnostic Systems, Sparks, MD, USA) and the resultant mixture was homogenized for 1–2 min, at low speed, in a sterile blender jar. The blended solution was then serially diluted from 10-1 to 10-8. Thereafter, 0.1 ml of each dilution was plated onto TSC agar plates and anaerobically incubated at 37 °C. After 24–48 h incubation, colonies showing morphology consistent with Clostridium isolates (i.e., black color) were selected for further testing to confirm their identity as C. perfringens as described above.

D. Screening for the presence of plc and cpe genes in C. perfringens isolates

Total C. perfringens DNA was isolated from the overnight TGY medium cultures, using the Wizard® Genomic DNA Purification Kit (Promega) and then subjected to PCR analysis using primers specific to each of the cpe and plc genes. The design of the primers was based on the C. perfringens strain SM101 genome sequence [8]. These PCR analyses utilized 100 ng template DNA, 25 pmol of each primer, 200 \(\mu\)M deoxynucleoside triphosphates (dNTPs) (Roche), 2.5 mM MgCl2, and 1 U Taq DNA polymerase (Fermentas) in a total volume of 50 \(\mu\)l. The reaction mixture was placed in a thermal cycler (Techne) for an initial period of 2 min at 94 °C, then 35 cycles, each 1 min at 94 °C, 1 min at 47 °C, 1 min at 72 °C, followed by an extension period of 10 min at 72 °C. The presence of a PCR amplified product was examined by subjecting an aliquot of each PCR sample to agarose (1.0%) gel electrophoresis, followed by ethidium bromide staining and photographing under UV light. The following primers were used to detect toxin genes:

Forward plc primer:

(5’-GATGGAAAAATTGATGGAACAGGAACT-3’),

Reverse plc primer:

(5’-CATGTAGTAGTCATCATCTGTTCCAGCATC-3’),

Forward cpe primer:

(5’-GGAGATGGTTGGTTGGATATTAGGGG-3’), and

Reverse cpe primer:

(5’-CTTCCAAGTCACATCTTTCGTCAG-3’)

E. DNA extractions, library preps, sequencing

Samples were prepared for sequencing with the Nextera DNA library prep. Twenty-five samples of C. perfringens isolates were multiplexed onto a single lane, 51 bp single end HiSeq3000 run, at the Center for Genome Research and Biocomputing (CGRB) at Oregon State University, Corvallis, OR, USA.

F. Bioinformatics: Assembly, Alignment to SM101 and genes, sequence logos of gene regions

Each sample was individually assembled with SPAdes [31]. SPAdes was run with default K-mers and the careful pipeline option to reduce mismatches and short indels. SPAdes assembly output scaffold files were compared against the C. perfringens SM101 genome [32], using nucmer [33], delta, and show-coords from MUMer [34]. Each sample was compared to eighteen C. perfringens genes as follows: sleC (ABG87393.1), virS (ABG86783.1), sigF (ABG87692.1), sigG (ABG86124.1), sigE (ABG85707.1), plc (ABG86694.1), cpe (ABG85760.1), cspB (ABG86463.1), gerKB (ABG85755.1), gerKA (ABG86956.1), gerKC (ABG86274.1), gerAA (ABG86934.1), spo0A (ABG85493.1), cpb (WP_003453250.1),

etx (WP_164789292.1), iap (WP_003463422.1), ibp (BAK40944.1) and netB (WP_110003253.1). BlastN [35] was used to identify the general genomic region of the gene in the assembly. When necessary, custom Perl scripts were used to extract and reverse-complement the gene sequence. Promer and show-snps from MUMer were used to identify translated amino acid single nucleotide polymorphisms (SNPs) between the assembled sample and the gene reference sequence. WebLogo [36] was used to evaluate the gene regions across all samples.

A. Isolation of C. perfringens

Our survey could detect C. perfringens isolates from all the targeted locations, i.e., the soil, floor, and food samples from selected hospitals (Table 1). Based on their characteristic, black colonies appearance on TSC plates (Table 1), the isolates were putatively identified as C. perfringens. The black colonies on TSC plate are a result of sulfite reduction by C. perfringens [37]. The soil samples obtained from all six hospital areas showed characteristic appearance of black colonies on TSC plate, albeit to varying degrees (Table 1). However, floor samples from only two of the six hospitals (Qassim hospital B and C) produced black colonies on TSC plate. Similarly, the food samples from only three of the six hospitals (Qassim hospitals A and B; Hail hospital D) produced the characteristic black colonies on the TSC plates (Table 1). When these isolates were subjected to Clostridium-specific biochemical tests, they fermented lactose, produced acid and gas, reduced nitrates to nitrites, and liquified gelatin within 48 h. Collectively, these results demonstrated that our survey successfully isolated Clostridium species from Hail and Qassim environments.

B. Differentiating C. perfringens from other Clostridium species by monitoring PLC and PFO phenotypes on blood agar plates

When newly-isolated Clostridium cultures were anaerobically grown on sheep blood agar plates, 25 out of 54 cultures produced a clear double (PFO-mediated inner and the \(\beta\)-toxin-mediated outer) zone of \(\alpha\)-hemolysis on the plates (Table 2). These results confirmed that these 25 cultures are C. perfringens isolates. However, the remaining 29 cultures, that were identified as Clostridium species by TSC plating and biochemical tests, failed to produce \(\alpha\)-hemolysis; no inner or outer zone of \(\alpha\)-hemolysis was detected on blood agar plates. These results suggest that other Clostridium species can be detected in soil, food, and hospital floor samples. However, further studies are required to support this hypothesis.

C. Detection of the plc and cpe genes in C. perfringens isolates by PCR

To confirm the presence of the plc gene in our newly-isolated C. perfringens, which produced PLC, we performed PCR analyses, using plc-specific primers. As a positive control, plc PCR analyses amplified the  900 bp plc-specific band from wild-type strains, SM101 and F4969. As expected from our blood agar plate phenotypic results, plc-specific PCR-amplified bands were obtained from DNA of all 25 surveyed isolates (Figure 1 A). As CPE is the main toxin responsible for C. perfringens type F FP, we examined whether our isolated C. perfringens carry cpe gene. Our PCR analyses, using cpe-specific primers, amplified the  1500 bp cpe-specific band from DNA of the reference strains SM101 and F4969. However, when DNA of our newly-isolated 25 strains were subjected to similar PCR analyses, no cpe-specific band was amplified from any DNA sample (Figure 1 B). Kuske et al. [38] had the same result; none of their surveyed C. perfringens soil isolates were cpe-positive. In contrast, Li et al [27] reported cpe-positive soil isolates harboring cpe gene on the plasmid. These findings support the possibility that soil may not be the major reservoir for C. perfringens chromosomal cpe isolates.

D. Comparing C. perfringens nucleotide sequences isolated from the KSA, USA, and Europe

We hypothesized that there might be differences in nucleotide sequence between the American/European and KSA strains, as FP outbreaks in the US and Europe are more frequent than those in the Middle East and KSA. To prove our hypothesis, we compared the whole genome sequence of Saudi ‘s C. perfringens strains with that of the American/European’s C. perfringens strains. SPAdes yielded de novo assemblies with a wide range of values; number of scaffolds ranging from 337–27220, assembly lengths ranging from 3.1–22.8 kb, and N50 ranging from 974–63344 bp (Table 3). When compared to the C. perfringens reference genome, most of the reference genome was found in the assemblies, excluding samples s007 and s008. On average, 87% of the C. perfringens genome was assembled in each sample. Samples s007 and s008 did not sequence well; s007 represented 53% of the C. perfringens genome and s008 only 9% (Table [t4]). Thirteen proteins were compared to the assembled samples. As previously mentioned, s007 and s008 were poor-quality assemblies that did not align well with the proteins. Sample s006 also showed a poor alignment with the protein sequences, along with many low-quality errors, due to low coverage in the assembly. Overall, the proteins and assembled contigs were well-matched, with only SNP differences. In the remaining 22 samples, the number of amino acid SNPs ranged from 0–89 (Table 1S), and the average across each gene was aligned to 22 genomes ranging from 0–44.5 SNPs (Table 4). cpe (ABG85760.1) was not found in any of the genome assemblies, which confirmed our PCR results (Figure 1B) that none of the 25 samples isolated from different places contained the cpe gene. gerAA (ABG86934.1) did not align with sample s001. plc (ABG86694.1) was aligned to all tested genome samples with average SNPs of 10.95 (Table 4). etx (WP_164789292.1) aligned to samples s019, s021, s022 only. Each KSA sample shared the same single SNP (Table 4). cpb (WP_003453250.1), iap (WP_003463422.1), ibp (BAK40944.1), and netB (WP_110003253.1) were not found when aligned to the KSA isolates.

Based on the data listed in Table [t6], proteins, that result from translating nucleotides with significant sequence differences between Saudi Arabian and American standard samples, may give rise to a similar protein pattern for both. For example, SigE (ABG85707.1) and SigF (ABG87692.1) have significantly different nucleotide sequences, but upon translation, there was no difference in protein patterns between the Saudi Arabian isolates and the American reference sample, SM101. Interestingly, all KSA strains, with the exception of sample numbers S23 and S24, had two proteins that differed from the reference sample SM101 in the SigF gene. SigE in all KSA strains showed no significant difference in amino acid sequence, indicating minimal change in protein function compared with that of SM101. Both nucleotide and protein sequences showed that the protein sequence was not affected by codons changing nucleotides (Table [t7]). This suggest that the underlying protein (and therefore protein function) is conserved while mutations occur. In other cases, the protein sequence did change, indicating that the amino acids at those positions were not conserved, which probably resulted in altered protein function (Table [t6]).

To determine the degree of protein conservation between the reference and KSA strains, we ran multiple alignments of a single region for some of the genes in C. perfringens that regulate toxin production and sporulation. Sequence logos were used for the multiple alignments of the protein sequences identified in each isolate for the following genes: gerKC (ABG86274.1) [39,40], spo0A (ABG85493.1) [41], plc (ABG86694.1) [42], and virS (ABG86783.1) [43]. Sequence logos showed a high sequence similarity between gene versions (Figure 2–5). We found that most Saudi Arabian samples had highly conservative proteins with the C. perfringens reference strain SM101, as shown in gerKC (ABG86274.1) (Figure 2), spo0A (ABG85493.1) (Figure 3), plc (ABG86694.1) (Figure 4), and virS (ABG86783.1) (Figure 5).