Bacteriological and molecular typing of Clostridium perfringens strains isolated in retail beef in Beijing, China

Abstract

Clostridium perfringens is an important zoonotic pathogen. This study was designed to explore the prevalence and toxin types of C. perfringens in retail beef collected from Beijing, China. Among 221 beef samples collected, 53 samples were positive for C. perfringens, resulting in the average prevalence as 23.98%. By toxin gene-based typing, the most C. perfringens strains belong to type A (96.23%, 51/53), only 2 strains were identified as type D. By a multi-locus sequence typing (MLST)-based analysis, a total of 36 sequence types (STs) were detected, and the most STs (n=30) represented just a single strain. These finding suggested that the prevalence of C. perfringens in retail beef in Beijing was considerably high and these bacteria displayed extreme diversity in genetics.

Clostridium perfringens is a well-known zoonotic pathogen in the world. It is widely distributed in nature and colonizes in the intestines of various animals, causing necrotizing enteritis, gas gangrene and sudden death [2, 3, 7, 16, 17]. Raw meat or meat products can be easily contaminated by this bacteria during the food processing, and subsequently cause severe illness in consumers [10, 19, 24].

In China, beef consumption displayed a gradual increase in past years, but few studies were conducted to explore how this meat was contaminated by C. perfringens. In this study, we conducted an investigation on the prevalence of C. perfringens in retail beef sold in Beijing, a metropolis consuming huge amount beef each day. As these beef came from a wide geographic ranges, the data obtained will be helpful to understand the overall prevalence of C. perfringens in China.

From 2019 to 2020, we collected a total of 221 fresh beef from 50 supermarkets in Beijing city. In laboratory, about 100 g of each beef sample was cut into small pieces and immersed in 100 ml sterile phosphate buffer saline (PBS) for one hour. The soaked solution was collected and centrifuged at 10,000 rpm for 5 min, then precipitation was transferred to C. perfringens selective chromogenic culture media plate (CHROMagar, Paris, France) and incubated in anaerobic jar with AnaeroGen (Oxoid Ltd., Basingstoke, UK) for 48 hr at 37°C. When colonies appeared on plate, 1–3 typical colonies in orange color were transferred to C. perfringengs agar base (Oxoid Ltd.) supplemented with SFP supplement (Oxoid Ltd.) for pure culture. Bacterial culture on the agar plate was washed off with sterile PBS solution and genomic DNA were extracted for species confirmation by PCR [23]. For each sample, only one isolate was used for following toxin gene-based typing and molecular typing analysis, according to the previously published PCR method [15, 20].

By bacterial culture, 53 samples were identified as positive for C. perfringens, leading to overall prevalence as 23.98% (Table 1). The beef marketing in Beijing came from four provinces, including Shandong, Hebei, Inner Mongolia and Shannxi. Relatively higher prevalence was found in the beef produced in Inner Mongolia (21/76, 27.63%), and the lowest prevalence was identified in the beef originated from Hebei province (14/72, 19.44%) (Fig. 1).

Table 1. Prevalence of Clostridium perfringens in retail beef with different geographic sources.

Geographic source	No. of samples collected	No. of samples positive for C. perfringens	Prevalence (%)	No. of type A C. perfringens	No. of type D C. perfringens
Shandong	65	16	24.62	16	0
Hebei	72	14	19.44	14	0
Inner mongolia	76	21	27.63	19	2
Shanxi	8	2	25	2	0
Total	221	53	23.98	51	2

Fig. 1.

Producing areas and prevalence of Clostridium perfringens-positive beef collected in this study.

By toxin gene-based typing, 53 typical C. perfringens strains were characterized into two toxigenic types, A and D (Supplementary Table 1). Type A was the main toxin type comprising of 51 strains (96.23%), which means that the majority C. perfringens strains harbor only cpa gene (producing alpha toxin). The remaining 2 strains (3.77%) were identified as type D, as cpa and etx genes (epsilon toxin) were detected in both strains. The cpb2 gene, which is responsible for production of beta 2 toxin (CPA2), was detected in 11 strains (20.75%), including two type D strains.

We used the multilocus sequence typing (MLST) scheme described by Jost et al. and Chalmers et al. to carry out phylogenetic analysis on these C. perfringens strains [3, 12]. Briefly, DNA fragments of eight housekeeping genes, including plc (cpa), ddlA, dut, glpK, gmk, recA, sod and tpiA, were amplified and sequenced at Sangon Biotech Co. (Sangon Biotech Co., Ltd., Shanghai, China). Eight sequences of each strain were analyzed and the allele numbers were assigned for different allelic sequences, then the sequence type (ST) was assigned for each strain using BioNumerics 8.0 software (Applied Maths, Keistraat, Belgium). Polymorphism of each locus were quantified using the Hunter & Gaston diversity index (HGDI) [9]. The minimum spanning tree (MST) was constructed using the MST method in BioNumerics 8.0 software. Clonal complexes were defined as groups of independent isolates that shared identical alleles at seven or more of the eight loci. The sequences determined for the MLST analysis have been deposited in the GenBank database under accession numbers MZ725945-MZ726368.

Among 53 C. perfringens strains, MLST typing identified 36 unique STs (ST1-36) (Supplementary Table 1). The ST14 contained the largest number of strains (n=8), 5 of them were originated from Inner Mongolia and 3 from Hebei province. The other frequently identified sequence type was ST32, containing 7 strains universally originated from Shandong province. Four STs (ST1, 16, 20 and 28) were comprised of just a couple of strains, while the remaining 30 STs were identified as singleton. High diversity was observed in the eight loci, with the HGDI value arranging from 0.660 to 0.921 (Table 2). The average number of alleles for all loci was 17. The loci plc and dut contained the largest number of alleles (n=22), while the minimum alleles was observed in recA and gmk (n=12). Unlike other loci, where the trimmed sequence from all the strains show an identical length, two different length fragments were detected in the dut gene (441 bp and 444 bp), with 3 alleles possessing a 3-bp internal insertion.

Table 2. Genes, number of alleles and Hunter & Gaston diversity index (HGDI) values of Clostridium perfringens strains obtained in Beijing (2019–2020).

Gene	Length (bp)	No. of alleles	HGDI
ddlA	429	18	0.854
dut	441 or 444	22	0.921
glpK	574	20	0.894
gmk	475	12	0.845
plc	544	22	0.794
recA	475	12	0.66
sod	478	17	0.697
tpi	451	13	0.844

MST analysis clustered 14 of 36 STs into 6 clonal complex (CC) subtypes, CC1 to CC6 (Fig. 2), the strains contained in the same CC were considered to be more closely related in genetics than other strains. The main subtype CC1 contained four STs (consisting of 11 strains), while the other subtypes (CC2-CC6) just contained two STs. In phylogenetic tree, there was no clear evidence that strains with the same geographic source likely clustered together. The two type D strains (S05 and S06) showed the closest genetic relationship and formed the CC6 in MST. While, the 11 strains harboring cpb2 gene almost evenly distributed across the MST tree (Supplementary Table 1, Fig. 2).

Fig. 2.

Minimal spanning tree constructed with the 53 Clostridium perfringens strains originated from retail beef in Beijing (2019–2020). Each circle represents a specific sequence type (ST), the strains sharing the same ST were included in the same circle. Different colors represent different geographic origins of these C. perfringens strains. The length between circles represents the genetic distance. The circles in shadow parts represents a clonal complex (CC), defined as group of independent strains that share identical alleles at seven or more of the eight loci. A total of six CCs (CC1-CC6) were identified among these group of strains. The strains representing ST 4 and 5 belong to type D C. perfringens. The 11 strains representing single ST respectively (including ST 2, 3, 4–7, 17, 18, 20, 21 and 26) harbors cpb2 in their genomes.

According to the data revealed in this study, the prevalence of C. perfringens in beef was around 24% in north area of China. This data is higher than the findings obtained in other countries. For example, in Seoul, Korea, the prevalence of C. perfringens in retail beef was below 16.67% [11]. While in Lahore city of Pakistan, prevalence of this organism in beef was no more than 1% [13]. In these beef-derived C. perfringens, type A was the most frequently observed toxin type. Although less frequently observed, etx-positive C. perfringens strains should be paid more attention as it can cause severe respiratory and neurologic signs in cattle [6]. On the other hand, cpb2 positive strains was also be found in apparently healthy animals [1, 21, 22], and its role in enteric disease of domestic animals and human needs to be further investigated.

Molecular typing schemes like MLST have been developed with the aim to associate C. perfringens strains to a specific disease presentation, toxin pattern or a specific host [3, 4, 8, 12, 18, 25]. However, in this study, no evidence was provided that a specific CC or ST could be highly associated with any toxin type, host sources or collection date. The lack of relevance between the phylogenetic features and biological properties may be due to the extremely open genome and genetic diversity observed in C. perfringens species [5, 14]. In the future, more conservative housing keeping genes even the whole genome based sequence should be used for correlation analysis between phylogenetic features and biological properties.

CONFLICTS OF INTEREST

The authors declare no conflict of interest.