Types A and F Clostridium perfringens in healthcare wastewater from Ghana

ABSTRACT

Clostridium perfringens can cause a plethora of diseases in humans and animals, making it a growing public health concern. Insight into C. perfringens genomic studies has enhanced our understanding of the diversity and evolution of this bacterium. However, most of these studies were centered on clinical strains from human and animal hosts, with limited data from environmental sources and developing countries. Here, we present genomic and phenotypic insights into C. perfringens strains from hospital wastewater samples from Ghana. Antibiotic susceptibility testing, toxinotyping, and whole-genome sequencing were performed on the seven C. perfringens isolates. The strains were susceptible to all antibiotics screened except for three strains that were resistant to clindamycin. The majority of these strains were type F strains, characterized by a chromosomally encoded cpe gene, whereas the remaining strains were classified as type A. The cpe gene is flanked upstream by IS1469 and downstream by IS1470. The genotype for the type F strains was ST721, whereas the type A strains were assigned to ST722, ST143, and ST211. Our type F strains clustered with other strains all within phylogroup I, which are mostly associated with foodborne illnesses, whereas the type A strains were grouped within phylogroup III. This study provides the first genomic characterization of C. perfringens from West Africa and contributes to further understanding of genomic diversity in C. perfringens. Moreover, the possible disseminated situation of C. perfringens in the environment highlights the need to monitor this bacterium in clinical specimens in Ghana.

IMPORTANCE

Clostridium perfringens causes gas gangrene and food poisoning in humans, and monitoring this bacterium is important for public health. Although whole-genome sequencing is useful to comprehensively understand the virulence, resistome, and global genetic relatedness of bacteria, limited genomic data from environmental sources and developing countries hamper our understanding of the richness of the intrinsic genomic diversity of this pathogen. Here, we successfully accumulated the genetic data on C. perfringens strains isolated from hospital effluent and provided the first evidence that predicted pathogenic C. perfringens may be disseminated in the clinical environment in Ghana. Our findings suggest the importance of risk assessment in the environment as well as the clinical setting to mitigate the potential outbreak of C. perfringens food poisoning in Ghana.

INTRODUCTION

Clostridium perfringens is an anaerobic, gram-positive bacterium that infects both humans and animals. It can also circulate in a wide variety of environments such as rivers and soil (1, 2). It possesses an arsenal of toxins responsible for disease pathogenesis, some of which serve as a framework for strain typing. Each strain is classified as toxinotype A–G based on the production of six key toxins: alpha-toxin (CPA), beta-toxin (CPB), epsilon-toxin (ETX), iota-toxin (ITX), enterotoxin (CPE), and necrotic enteritis B-like toxin (NetB) (3).

Recent insights from whole-genome sequencing (WGS) analysis of C. perfringens have enhanced our understanding of the genetic variation and evolution of this pathogen (4, 5). Comparative analysis of C. perfringens and other bacterial species has revealed high genomic openness and genetic diversity of C. perfringens, which could indicate a significant horizontal gene transfer event facilitating their adaptation to different niches (6, 7). The use of WGS has also led to accurate genomic characterization and the discovery of novel determinants of C. perfringens. For instance, the discovery of netB, associated with avian necrotic enteritis and becA/B, was investigated using genomic sequencing (8, 9).

C. perfringens is grouped into five (I–V) main phylogenetic lineages based on its core genomes (7). This grouping contradicts the clustering based on C. perfringens accessory genes, which are grouped into three main lineages. However, there seems to be an association between the core genome-based clustering and accessory gene-based clustering of strains within phylogroups I and II, but not with strains within phylogroup III–V (7). The utilization of C. perfringens genomic data from several niches and geographical spaces could further help to understand the association between the core genome and accessory gene-based clustering (7). However, most of the available genomes of C. perfringens are derived from strains isolated from humans, animals, and food products, whereas those from environments such as soils and wastewater are comparatively few (10). Especially, hospital wastewater poses a substantial threat to human health and the ecological environment due to its potential harboring of pathogenic and antibiotic-resistant bacteria, leading to contamination of surrounding areas, including aquatic ecosystems (11–13). Despite the significance thereof in monitoring the dissemination of clinical virulent pathogens, including toxin-producing C. perfringens in hospital wastewater (11–13), data involved in the isolation and genomic characterization of C. perfringens have not been accumulated particularly from West African region. This situation should be improved because it can help with the surveillance of local C. perfringens distribution (14).

Thus, genome collection from unexplored environmental sites, such as hospital wastewater in Ghana, may strengthen our understanding of the genetic characteristics of C. perfringens as well as monitoring the basis for efflux of C. perfringens from clinical environments. In this study, we aimed to explore the antibiotic susceptibility profile and genetic features of C. perfringens isolates from hospital wastewater samples in Ghana for the first time, where extremely limited information on C. perfringens exists.

RESULTS

Strain identification and PCR screening of virulence factor genes

To confirm the species of each of the seven non-duplicate tank-derived strains with positive for the lecithinase activity, Sanger sequencing of the 16S rRNA gene was conducted. All strains were classified into C. perfringens with more than 99.9% of the identities.

Subsequently, we conducted multiplex PCR to investigate the presence of virulence factors like plc, cpe, and becA/B found in human enteropathogenic strains (15). All strains harbored plc-encoding CPA, which is involved in the pathogenicity of C. perfringens gas gangrene (16). Four strains had cpe, whereas becA/B, a novel binary toxin gene recently reported in C. perfringens (9), was absent in all strains (Fig. 1).

Fig 1.

Multiplex PCR detection of plc, cpe, and becA/B in C. perfringens strains isolated in Ghana. A 100-bp ladder (Biomedical Science, Tokyo, Japan) was used as a molecular marker. The positive-band sizes of each gene are as follows: plc, 324 bp; cpe, 233 bp; becA, 499 bp; and becB, 416 bp.

Antibiotic susceptibility profile

To determine the antibiotic susceptibility profile of each strain, the disk diffusion method was conducted using six antibiotics as described in the European Committee on Antimicrobial Susceptibility Testing v. 13.0 (https://www.eucast.org/fileadmin/src/media/PDFs/EUCAST_files/Breakpoint_tables/v_13.0_Breakpoint_Tables.pdf). C. perfringens strains were highly susceptible to most of the antibiotics used for the treatment of clostridial sepsis. All seven strains were susceptible to all antibiotics screened, except for clindamycin. Of the seven strains, ENRHS10, 14, and 16 exhibited clindamycin resistance, and the remaining strains were susceptible to clindamycin (Table 1).

TABLE 1.

Antibiotic susceptibility profiles of C. perfringens strains isolated in Ghana

Antibiotic	Interpretive category of zone diameter
	ENRHS1	ENRHS3	ENRHS4	ENRHS7	ENRHS10	ENRHS14	ENRHS16
Penicillin G	Sa	S	S	S	S	S	S
Penicillin/tazobactam	S	S	S	S	S	S	S
Meropenem	S	S	S	S	S	S	S
Vancomycin	S	S	S	S	S	S	S
Clindamycin	S	S	S	S	R	R	R
Metronidazole	S	S	S	S	S	S	S

^a S, susceptible; R, resistant.

Genomic characterization of C. perfringens strains

We further characterized the genomic feature of each isolate using WGS, and complete genomes of all were successfully constructed using Unicycler ver. 0.5.0 (17). The genomes of these C. perfringens strains ranged from 3.0 to 3.3 Mb (Table S1). All genomes, except ENRHS16, had a circularized chromosome and one extrachromosomal element. Four strains (ENRHS1, ENRHS3, ENRHS4, and ENRHS7) had almost the same chromosomal size of 3.0 Mb and an extrachromosomal element of 13,729 bp. The remaining three strains, ENRHS10, ENRHS14, and ENRHS16, had the largest genomes. For instance, the genome of ENRHS16 comprises a chromosome and two other circular plasmids of 47,031 and 59,543 bp in size (data not shown). The number of coding sequences recorded among these C. perfringens genomes ranged from 2,711 to 2,952 (Table S1).

Moreover, to investigate the relatedness of strains ENRHS1, ENRHS3, ENRHS4, and ENRHS7 with similar chromosomal sizes using multilocus sequence typing (MLST), we found that all the four strains were assigned to sequence type (ST) 721 (Table S2). In contrast, strains ENRHS10, ENRHS14, and ENRHS16 exhibited distinct STs, specifically ST143, ST722, and ST211, respectively (Table S2).

Phylogenetic analysis

From core gene alignment using Roary ver. 3.13.0 (18), 10 core genes were identified, and 2,730–2,733 bp of DNA sequences were aligned. The ModelFinder (19) in IqTree ver. 2.2.0.3 (20) determined TIM2+F+I as a best-fit model according to the Bayesian information criterion. C. perfringens genome was revealed to be classified into five phylogroups, some of which were associated with specific pathogenetic strains (7). Notably, Phylogroup I is linked to enteropathogenic strains harboring cpe typically isolated from food poisoning (FP) cases (7). Thus, to infer the comprehensive genetic relationship between our strains and each phylogroup, the phylogeny of the seven C. perfringens from Ghana, 18 reference genomes (7) representing each phylogroup, and Clostridium massiliamazoniense as an outgroup (21), was investigated. The resultant core-genome-based phylogenetic tree categorized these 25 C. perfringens strains into the five phylogroups as previously described (7). Notably, all type F strains from this study clustered within phylogroup I, where the majority of enteropathogenic strains associated with foodborne illnesses belong. In contrast, type A strains were in phylogroup III, which is mostly heterogeneous in nature.

Virulome and resistome analysis

First, to characterize major toxin prevalence among Ghanaian strains, toxinotyping was conducted (3). Of the seven toxinotypes currently defined, only types A and F were detected among our strains. Notably, Toxinotype F was the most predominant (4/7, 57.1%), and strains belonging to this type were characterized by cpe, a toxin gene known to cause gastrointestinal diseases (22, 23) (Fig. 2).

Fig 2.

Maximum-likelihood phylogenetic analysis based on C. perfringens core-genomes. The tree includes 7 Ghanaian strains in this study, 18 reference genomes, and C. massiliamazoniense as an outgroup (21). The tree was constructed using IqTree ver. 2.2.0.3 (20). The rectangles colored in five colors suggest phylogroup where each node belongs. Toxinotype is shown in the location of the navy circle. Profile of virulence genes was searched using Virulence Factor Database (VFDB) (24) and local-BLAST. AMR means antimicrobial resistance. AMR genes were detected by ABRicate. The red box means the presence of virulence genes, and the green box means the presence of AMR genes. White box means the absence of genes.

Furthermore, in our investigation of other virulence factors contributing to bacterial pathogenicity, we employed the Virulence Factor Database (VFDB, http://www.mgc.ac.cn/cgi-bin/VFs/v5/main.cgi) (24). The four type F isolates showed similar profiles for the presence of other virulence genes (Fig. 2). Notably, these strains lacked the nan and nag genes, which are widely prevalent in C. perfringens, and encode sialidases and hyaluronidases, respectively (Fig. 2). However, the virulence characteristics of the three type A strains varied. Specifically, cpb2 encoding beta2 toxin was present in ENRHS10 and ENRHS16, but not in ENRHS14, whereas pfoA encoding perfringolysin O (PFO) was present in only ENRHS14 and ENRHS16. The gene encoding bile salt hydrolase (BSH) bsh was present in only ENRHS16 (Fig. 2).

Given the previous reports of multidrug-resistant C. perfringens isolates in clinical and retail samples (25, 26), we conducted resistome analysis using ABRicate ver. 1.0.1 (https://github.com/tseemann/abricate) against multiple database like the comprehensive antibiotic resistance database (CARD) (27), Resfinder database (28), NCBI bacterial antimicrobial resistance reference gene database (29), and MEGARes 2.0 database (30). The results of our analysis unveiled the presence of tetA in all strains, and tetB was limited to ENRHS10 and ENRHS16 (Fig. 2).

Comparative analysis of the genetic organization of cpe loci

The chromosomal cpe was considered as a conserved region, whereas plasmid-mediated cpe exhibited a variety of sequence patterns (31, 32). However, recent large-scale analyses of pathogenic strains involved in foodborne illness outbreaks have shown rearrangements in the order of closely located genes, resulting in diverse cpe surrounding genetic structures (33).

To explore the presence of genetic diversity in cpe loci among Ghanaian strains, we conducted a comparative genetic structure analysis of cpe loci extracted from WGS data, utilizing EasyFig ver. 2.2.5 (34). Our analysis revealed that all cpe loci exhibited identical genetic organization (Fig. 3). Comparative analysis of cpe loci between Ghanaian strains and NCTC 8239, the historical chromosomal cpe positive strain, revealed that only the Ghanaian strains had the downstream IS1470 gene adjacent to cpe (Fig. 3). This particular genetic structure had not been reported in previous research (33). The comparative analysis of cpe loci highlights the potential for greater diversity in the genetic environment and arrangement of cpe loci than initially observed (31–33), offering a valuable tool for the sub-classification of cpe-positive strains.

Fig 3.

Comparison of the genetic environment around cpe locus. Comparison of the genetic environment around cpe locus between four type F strains and reference genome of chromosomal cpe-positive strain, NCTC 8239, was visualized by Easyfig ver. 2.2.5 (34). Genome data of NCTC 8239 were downloaded from NCBI database (PRJEB6403).

DISCUSSION

C. perfringens has the potential to cause diverse pathogenicity in humans and animals because of the plethora of virulence factors. However, this opportunistic pathogen is primarily acquired by specific populations of human infants (35). These infections include gas gangrene, FP, antibiotic-associated diarrhea, hemorrhagic enteritis, and necrotic enteritis (36). Utilizing accumulated WGS data of C. perfringens can provide meaningful insights into the genomic diversity and phylogenetic characteristics of the species (4–7). However, according to the PubMLST database, strong geographic bias is still observed as all available genomes from the African region are derived from Egypt or South Africa. To further explore the genomic openness and genetic diversity of C. perfringens (6, 10), WGS of isolates, particularly from regions such as West Africa, where such data are limited, are needed. Here, we obtained and characterized seven complete genomes of C. perfringens isolates from clinical wastewater samples in Ghana for the first time. Resistome analysis revealed the presence of several tetracycline resistance genes, including tetA in all isolates and tetB specifically in ENRHS10 and ENRHS16. Notably, MLST analysis suggested that all type F strains from different sewage tanks were newly defined as ST721. Focusing on the genetic environment around the cpe locus revealed that the genetic structure surrounding cpe in Ghanaian strains was different from that of NCTC 8239, a historical chromosomal cpe-positive strain, which could imply the genetic diversity of cpe locus and lead to further sub-classification of chromosomal cpe positive strains.

Of the seven strains from Ghana, four isolates harbored cpe and belonged to toxinotype F. Type F strains are typically involved in enteric diseases, such as FP and antibiotic-associated diarrhea, whereas previous studies reported that type F C. perfringens was isolated from fecal samples of healthy adults (37, 38). All four type F strains clustered with the other strains in phylogroup I (Fig. 2), suggesting that Ghanaian type F strains, may possess the pathogenic potential to humans. In addition, these strains usually cause FP, and the majority lacks the nanI and nanJ genes but harbor the nanH gene (7, 39). These genes encode sialidase, which catalyzes the removal of free sialic acid from glycolipids and glycoproteins; however, the contribution of sialidase to the virulence of C. perfringens remains unclear (39). According to Abdel-Glil et al., four hyaluronidase genes (nagI, -J, -K, and -L) are absent, and a truncated nagH is present in most phylogroup I strains (7). The Ghanaian type F strains also showed the same genetic variation, suggesting that the contribution of hyaluronidase to the pathogenicity of phylogroup I strains could be common. Hyaluronidase is non-toxic and plays a role in facilitating the spread of α-toxin, a known cause of gas gangrene (40). However, the absence and truncation of nag genes in phylogroup I could suggest that they are less important for the pathogenicity in phylogroup I strains. The virulence nature (prevalence of toxin and hyaluronidase genes) of the type F strains from the Ghanaian hospital sewage samples should heighten local public health concerns because they could have resulted in toxico-infection, which might have gone undetected or have had the etiological agent incorrectly assigned. The spread of such strains poses a public health threat; hence, the microbiological examination of suspected cases in the study area should be encouraged. Infections with type A strains on the body surface lesions can cause gas gangrene and myonecrosis. The phospholipase C enzyme, also referred to as CPA, is thought to be the main virulence factor for type A strains; however, there is the possibility of type A-mediated enteric disease pathogenicity by unknown virulence factors (41). This is based on the fact that the present type G strains with NetB toxin, which causes necrotic enteritis in chickens, were once part of type A strains (8, 36). In addition, a genomic study of FP-related strains from outbreaks in France reported a number of cpe-negative strains (33).

In this study, the C. perfringens genomes of type F strains had a relatively small genome size, as was previously described (7). All three type A isolates clustered within phylogroup III (Fig. 2) and exhibited diverse genomic characteristics. Virulence factors that have previously been described in type A strains include pfoA, bsh, and cpb2, and sialidase and hyaluronidase genes have been described in type A strains. Only two of our type A strains (ENRHS14 and ENRHS16) harbored pfoA coding for PFO. PFO and α-toxin function synergistically, contributing to the development of gas gangrene (42). Recent research has established an association between the presence of pfoA and cytotoxicity to human gut epithelial cells in vitro, as well as in vivo using a murine model (43). The presence of pfoA in ENRHS14 and ENRHS16 indicates a heightened virulence potential, raising concerns regarding their contamination of hospital wastewater that is released into the external environment. This contamination poses an increased risk to public health.

ENRHS16 harbors a plasmid-mediated bsh gene that codes for BSH. BSH plays a pivotal role in catalyzing the hydrolysis of conjugated bile acids into unconjugated bile acids (44). In C. perfringens, conjugated bile salts have been recognized for their role in spore germination (45). However, despite its identification and well-established enzymatic characterization several decades ago (46, 47), the specific role of BSH in pathogenicity and host colonization in C. perfringens has remained unexplored. Previous research in other bacterial species suggests that BSH may contribute to host colonization (48, 49). In Brucella abortus, BSH contributes not only to growth in the presence of bile acid but also to infectivity in mice (48). In Listeria monocytogenes, the bsh-deficient strain showed a decreased bile acid persistence and lower survival rate in the intestinal tract (49). These previous findings suggest that BSH may support bacterial colonization to host intestinal tract in C. perfringens, which also enables host-mediated environmental dissemination, although further experiment is required to validate this prediction.

Although genomic data of environmental C. perfringens strains have improved our understanding of this species (10, 50), to the best of our knowledge, the genomic characteristics of C. perfringens in clinical wastewater have not been explored. In addition, in Ghana, an outbreak of C. perfringens at a senior high school has been observed, stressing the public health importance of C. perfringens (51). Despite this, a complete genome of the Ghanaian isolates was not obtained prior to this study. This is the first genomic study of C. perfringens genome collection in Ghana and can contribute to local infection surveillance. This research also contributes to the enrichment of the genomic database of C. perfringens and encourages worldwide genomic study of this bacterium. Our findings revealed that newly defined ST lineages may contribute to a further deep understanding of the genomic characteristics including evolution among C. perfringens lineages, leading to spectating the comprehensive biological role of unknown functional genes.

This study had some limitations. Only seven C. perfringens isolates were characterized, which do not necessarily represent the characteristics of all Ghanaian strains. To conduct a comparative genome analysis to reveal genomic features of other Ghanaian strains, more genomic data need to be collected, which will be the next challenge for our future research plan. Moreover, we could not measure C. perfringens infection cases in the hospital and its surrounding region due to the lack of a surveillance system.

In conclusion, this study identified the contamination of the predicted pathogenic C. perfringens in clinical wastewater in Ghana and provides the first genomic basis. This genomic characterization study from West Africa may contribute to advancement not only for further understanding the genomic diversity in C. perfringens but also to strengthen public health surveillance of this bacterium worldwide. Moreover, our findings stress the need to monitor toxin-producing strains in clinical specimens, leading to appropriate infection control measures for this bacterium in Ghana.

MATERIALS AND METHODS

Isolation and identification of C. perfringens

Hospital sewage samples were collected from 16 different sewage tanks which are independently located sites at the Effia Nkwanta Regional Hospital in the western region of Ghana. Each sample was individually inoculated in 17 mL thioglycolate medium (Becton Dickinson, Franklin Lakes, NJ) and incubated at 37°C for 3 days. Aliquots of each growth culture were plated on Clostridium welchii agar plates containing kanamycin (Nissui, Tokyo, Japan) and incubated at 37°C under anaerobic conditions. Colonies positive for the lecithinase activity were anaerobically passaged on 5% sheep blood agar plates. Anaerobic condition was achieved using AnaeroPack Kenki and AnaeroPack jar (Mitsubishi Gas Chemical, Tokyo, Japan).

16S rRNA Sanger sequencing

Genomic DNA was extracted from selected colonies using the Cica Geneus DNA extraction reagent (Kanto Chemical, Tokyo, Japan). For the amplification of 16S rRNA gene from each DNA sample, we employed the primer set (Table S3). The PCR reaction mix followed this protocol: an initial denaturation step at 98°C for 1 min, followed by 30 cycles of 98°C for 5 s, 57°C for 10 s, and 72°C for 1 min, with a final extension at 72°C for 3 min (52). Subsequently, each PCR product was purified using the ExoSAP-IT PCR Product Cleanup Reagent (Thermo Fisher Scientific, Waltham, MA) in accordance with the manufacturer’s protocol. The purified amplicons were sequenced using a 3730xl DNA Analyzer (Thermo Fisher Scientific) utilizing the PCR primer set as sequencing primers. Bacterial species were identified by comparing the obtained 16S rRNA gene sequences with the top-ranked species in the 16S rRNA database (https://blast.ncbi.nlm.nih.gov/Blast.cgi). The primer sets used for the 16S rRNA gene amplification and sequencing are listed in Table S3.

Multiplex PCR for detecting C. perfringens toxin genes

The presence of C. perfringens toxin genes such as plc, cpe, and becA/B causing enteropathogenic effects was screened using multiplex PCR as previously described (15). The primer sets used are listed in Table S3.

Antibiotic susceptibility testing

Susceptibility of the strains to vancomycin, clindamycin, metronidazole, penicillin G, piperacillin/tazobactam, and meropenem (Mast Group, Bootle, UK) was evaluated using the disk diffusion method based on the European Committee on Antimicrobial Susceptibility Testing v. 13.0 (https://www.eucast.org/fileadmin/src/media/PDFs/EUCAST_files/Breakpoint_tables/v_13.0_Breakpoint_Tables.pdf). Briefly, overnight cultures were adjusted to the turbidity of McFarland no. 1 and plated onto Fastidious Anaerobe Agar plates (NEOGEN, Lansing, MI) containing 5% horse blood (Nippon Bio-Supp. Center, Tokyo, Japan) and incubated for 20 h under anaerobic conditions. Zones of inhibition were interpreted using the criteria of the European Committee on Antimicrobial Susceptibility Testing v. 13.0. The testing was performed with a single measurement for all antibiotics.

WGS and hybrid de novo assembly

To fully circularize the genomes of these C. perfringens strains and accurately determine their genetic features, short- and long-read sequencing were conducted for each isolate. Genomic DNA was extracted using the NucleoSpin Tissue Kit (Takara Bio, Shiga, Japan) for short-read sequencing. The sequencing library was prepared using Illumina DNA Prep and sequenced using Mid Output Reagent Cartridge (300 cycles) on Illumina Miniseq platform (Illumina, San Diego, CA). For long-read sequencing, genomic DNA was extracted using the MagAttract HMW DNA Kit (QIAGEN, Hilden, Germany), and the DNA library was prepared with Native Barcoding Expansion 1–12 (EXP-NBD104) and ligation sequence Kit (SQK-LSK109). Sequencing was performed using a MinION flow cell FLO-MIN106 R9.41 in a MinION sequencer (Oxford Nanopore Technologies, Oxford, UK). Short reads with low-quality scores (reads with Q < 30) and short lengths (reads < 10 bp) were filtered using Fastp ver. 0.12.4. Raw data from long-read sequencing were obtained in fast5 file format and were subjected to base-calling with Guppy ver. 6.1.5. Low-quality scores (reads with Q < 10) and short lengths (reads < 1,000 bp) included in output fastq files were filtered using Filtlong ver. 0.2.1. A hybrid assembly of filtered short and long reads was performed using Unicycler ver. 0.5.0 (17).

In silico analyses

The MLST tool, available at the Center of Genomic Epidemiology (https://cge.food.dtu.dk/services/MLST/), was used to identify the STs of the isolates. To classify the strains based on their toxin profiles, toxinotyping was conducted using the local BLAST tool with a cut-off value of 70% identity and 30% coverage. A query containing amino acid sequences of seven typing toxins (plc, cpe, cpb, netB, etx, and iap/ibp) collected from UniProt was performed against each genome. The presence of other virulence factors, such as sialidases and hyaluronidases, was determined using the local BLAST and VFanalyzer tool from the VFDB (24). The presence of antibiotic resistance genes in these genomes was identified using ABRicate against CARD (27), Resfinder (28), NCBI bacterial antimicrobial resistance reference gene database (29), and MEGARes 2.0 database (30) with a cut-off value of 90% identity and 30% coverage.

Phylogenetic analysis

To describe the phylogeny of these C. perfringens isolates and establish their genomic characteristics, reference genomes of C. perfringens data belonging to each phylogroup, as published by Abdel-Glil et al. (7), and other sequences were collected from the NCBI database. As an outgroup, C. massiliamazoniense was included (21). All genomes (7 complete genomes from this study, 18 draft genomes, and the genome of C. massiliamazoniense) were annotated using Prokka ver. 1.14.6 (53). Roary ver. 3.13.0 (18) was used to construct a pan-genome, and a core-genome alignment file was obtained. This file was inputted into IqTree ver. 2.2.0.3 (20) to generate a phylogenetic tree using maximum-likelihood method. The best-fit model was determined using ModelFinder (19). The tree was visualized and annotated with the toxinotype, antibiotic resistance, and virulence genes using iTOL (https://itol.embl.de/) (54).

Comparison of genetic environments around cpe locus between Ghanaian strains and NCTC 8239

Although the genetic organization of cpe loci was considered to be a conserved region among strains (31, 32), a recent genomic study of C. perfringens involved in foodborne outbreaks reported diverse genetic structures around cpe genes (33). To investigate the genetic variation of cpe loci in Ghanaian strains, we conducted a comparative genetic analysis of the cpe locus among the newly characterized genomes of Ghanaian strains. This comparative analysis was facilitated using Easyfig ver. 2.2.5 (34). Furthermore, to characterize their cpe loci and make comparisons with other representative cpe loci, we used NCTC 8239, a well-known historical chromosomal cpe-positive strain, as a reference for our analysis (14).

DATA AVAILABILITY

All sequence data and related information of seven Ghanaian isolates are available at National Center for Biotechnology Information with Biosample accession numbers: SAMN33819949, SAMN33820436, SAMN33824391, SAMN33824413, SAMN33826306, SAMN33826312, and SAMN33826337. Reference genome to identify phylogroup is obtained following Abdel-Glil M et al. (7). NCTC 8239 sequencing data were obtained from Bioproject PRJEB6403 (ENA accession number: SAMEA4063013). C. massiliamazoniense genome used in phylogenetic analysis was obtained from Bioproject PRJEB4195 (accession number: SAMEA3538361.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aem.01619-23.

Tables S1 to S4. aem.01619-23-s0001.pdf.

Characterization of Type A and F C. perfringens used in this study.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.