Specific PCR Identification and Differentiation of the Thermophilic Campylobacters, Campylobacter jejuni, C. coli, C. lari, and C. upsaliensis

Abstract

A sensitive PCR assay that detects the thermophilic campylobacters C. jejuni, C. coli, C. lari, and C. upsaliensis is reported. Furthermore, by digestion of the PCR products with two restriction enzymes, species differentiation was demonstrated. Thus, the present method has the potential to be used for both detection and identification of thermophilic Campylobacter species.

The rate of Campylobacter infections has been increasing, with the number of cases often exceeding those of Salmonella and Shigella (14, 16). The clinically most important campylobacters are the members of the thermophilic group, C. jejuni, C. coli, C. lari, and C. upsaliensis, with C. jejuni responsible for the majority of human cases (1). The true incidences, however, might be underestimated, since most cultivation methods are designed for C. jejuni alone and therefore may not be optimal for the other species (3). Furthermore, in many clinical laboratories, typing is only done to the genus level, since species determination is cumbersome and sometimes gives incorrect or intermediate results.

Thus, there is a great need for simple methods for detection and reliable differentiation of the thermophilic Campylobacter species. Indeed, a number of genetically based detection and typing methods have been developed for Campylobacter species in food, clinical, and environmental samples (4–6, 9, 12, 13, 17, 22, 23). However, the majority of these methods suffer from drawbacks like the need for multiple PCRs or restriction enzymes or the requirement of a hybridization step with species-specific probes after the PCR.

This study describes a sensitive and simple method for both the detection and differentiation of thermophilic campylobacters. Detection is achieved by the amplification of 491 bp of a highly polymorphic part of the 23S rRNA gene, and further species differentiation is accomplished by digestion of the PCR product by two restriction enzymes, AluI and Tsp509I, resulting in specific restriction fragments for each species.

The bacterial strains used in this study are listed in Table 1. The bacteria were cultured aerobically or in a microaerophilic atmosphere on blood agar at 37 or 42°C.

TABLE 1.

Bacterial strains used in this study

Species	No. of strains tested	No. of PCR-positive strains	Species according to PCR typing	Sourcea
C. jejuni
Human isolates	66	66	65 C. jejuni, 1 C. colib	UU, RSÖ, Telelab
Veterinary isolates	7	7	C. jejuni	NVI
SLV-167	1	1	C. jejuni	NFA
SLV-267	1	1	C. jejuni	NFA
SLV-264	1	1	C. jejuni	NFA
CCUG 11284	1	1	C. jejuni	CCUG
C. coli
Human isolate	1	1	C. coli	Telelab
Veterinary isolates	14	14	13 C. coli, 1 C. jejunib	NVI
SLV-271	1	1	C. coli	NFA
CCUG 11283	1	1	C. coli	CCUG
C. lari
Veterinary isolates	3	3	C. lari	NVI
CCUG 23947	1	1	C. lari	CCUG
From seagulls	2	2	C. lari	NVI
4219/92	1	1	C. lari	NIPH
1557/93	1	1	C. lari	NIPH
1089/94	1	1	C. lari	NIPH
2138/94	1	1	C. lari	NIPH
C. upsaliensis
Human isolate	1	1	C. upsaliensis	Telelab
Veterinary isolates	6	6	C. upsaliensis	NVI
From dogs	5	5	C. upsaliensis	NVI
CCUG 14913	1	1	C. upsaliensis	CCUG
14913	1	1	C. upsaliensis	NIPH
C. hyointestinalis	2	0		NVI
C. fetus subsp. fetus	2	0		NVI
C. fetus subsp. venerealis	1	0		NVI
C. mucosalis	2	1c		NVI
Helicobacter pylori	8	0		UU
Salmonella enteritidis SLV-397	1	0		NFA
Salmonella typhimurium SLV-248	1	0		NFA
Salmonella dublin SLV-242	1	0		NVI
Salmonella bovismorbificans SLV-443	1	0		SDIIC
Shigella flexneri 820/92	1	0		SDIIC
Shigella boydii 915/94	1	0		SDIIC
Shigella sonnei 637/94	1	0		SDIIC
Escherichia coli
SLV-165	1	0		NFA
O157:H7 81186	1	0		SDIIC
O157:H7 ATCC 43889	1	0		SDIIC
O157:H7 ATCC 43890	1	0		SDIIC
Yersinia enterocolitica SLV-408	1	0		NFA
Vibrio cholerae CCUG 34649	1	0		CCUG
Vibrio parahaemolyticus CCUG 19113	1	0		CCUG
Vibrio vulnificus CCUG 13448	1	0		CCUG
Staphylococcus aureus SLV-350	1	0		NFA
Staphylococcus saprophyticus SLV-013	1	0		NFA
Enterobacter cloacae SLV-164	1	0		NFA
Enterococcus durans SLV-078	1	0		NFA
Bacillus cereus SLV-160	1	0		NFA
Klebsiella oxytoca SLV-089	1	0		NFA
Clostridium perfringens SLV-442	1	0		NFA

^aNFA, National Food Administration, Uppsala, Sweden; SDIIC, Swedish Institute for Infectious Disease Control, Stockholm, Sweden; NVI, National Veterinary Institute, Uppsala, Sweden; Telelab, A/S Telelab, Skien, Norway; UU, University of Uppsala, Uppsala, Sweden; RSÖ, Örebro Medical Center Hospital, Örebro, Sweden; CCUG, Culture Collection of the University of Göteborg, Göteborg, Sweden; NIPH, National Institute of Public Health, Oslo, Norway. 

^bThe disagreements between the biochemical and genetic typing are discussed in the text. 

^cOne strain of C. mucosalis gave a faint band of the correct size when using 1.5 mM MgCl₂. This band disappeared when the MgCl₂ concentration was lowered to 0.5 mM. 

Chromosomal DNA was prepared either by simply breaking the bacterial cells with heat or by the method of Pitcher et al. (15). The PCR mixtures contained a 0.25 μM concentration of each oligonucleotide (Scandinavian Gene Synthesis, Köping, Sweden), 50 mM KCl, 10 mM Tris-HCl (pH 8.3 at room temperature), 1.5 mM MgCl₂, a 0.1 mM concentration of each deoxynucleotide, and 0.02 U of AmpliTaq Gold (PE Applied Biosystems, Branchburg, N.J.) per μl. The temperature cycling was performed on a PTC-200 DNA engine (MJ Research, Watertown, Mass.) at 94°C for 1 min, 56°C for 1 min, and 72°C for 1 min for 45 cycles, followed by a final extension at 72°C for 5 min. The temperature protocol was preceded by a 12-min incubation at 94°C in order to activate the AmpliTaq Gold polymerase. The PCR products were visualized by 3% agarose gel electrophoresis. To differentiate between species, 10-μl samples were removed from PCR-positive mixtures and digested in two separate reaction tubes for 1 h with either 0.5 U of AluI at 37°C or 0.5 U of Tsp509I at 65°C. PCR products of type strains with known banding patterns were digested and analyzed in parallel with unknown samples in order to ensure cleavage activity. Banding patterns, easy to interpret, were reproducibly obtained and visualized by gel electrophoresis (Fig. 1).

FIG. 1.

Restriction fragments generated by digestion with AluI and Tsp509I of PCR products of the thermophilic campylobacters. Lanes: 1, C. jejuni CCUG 11284; 2, C. lari CCUG 23947; 3, C. coli CCUG 11283; 4, C. upsaliensis C434 (from a dog); 5, C. upsaliensis CCUG 14913. Lanes M contain a 100-bp ladder.

The most variable part of the 23S rRNA gene, located between helices 43 and 69 (7), of different Campylobacter species has been extensively sequenced and used as a target for species-specific PCR methods (4, 8). One primer pair, THERM1 and THERM2, has been shown to specifically detect the thermophilic Campylobacter species (4). In the present study, THERM1 was used together with a new primer, THERM4 (5′-CTTCGCTAATGCTAACCC-3′). THERM4 was constructed to perfectly match the sequences of the thermophilic species by using published DNA sequences (Table 2). The primer pair THERM1 and THERM4 demonstrated high specificity, giving a PCR product of the correct size for all strains of the thermophilic group, but not for other Campylobacter species or other bacteria (Table 1). One isolate of Campylobacter mucosalis, though, gave a faint band of the correct size. This band, however, disappeared when the MgCl₂ concentration was lowered to 0.5 mM. Nevertheless, the higher MgCl₂ concentration was used throughout the study, since amplification of the C. mucosalis isolate resulted in a barely visible band on the gel and the restriction fragments of this PCR product generated unique banding patterns clearly distinguishable from those of the target species.

TABLE 2.

23S rRNA gene sequences accessible in the GenBank sequence database used for primer design and for choosing discriminatory restriction enzymes

Strain	Accession no.	Source or reference
C. jejuni
LMG 6444	X67767	21
LMG 6629	X67765	21
ATCC 33250	X66616	20
LMG 7790	X67766	21
ATCC 43431	Z29326	11
C. coli
LMG 7535	X67770	21
LMG 6440	X67764	21
VC167	U09611	19
C. lari
LMG 8846	X67769	21
NCTC 11352	Y11764	Direct submission
C. hyoilei NCTC 33450	Y11765	Direct submission
C. upsaliensis
LMG 8854	X67774	21
LMG 7533	X67763	21
C. sputorum subsp. sputorum LMG 7795	X67758	21
C. sputorum subsp. fecalis LMG 6617	X67758	21
C. sputorum subsp. bubulus LMG 6447	X67772	21
C. mucosalis
LMG 6448T	X80636	2
LMG 7794	X67760	21
C. fetus subsp. venerealis LMG 6570	X67773	21
C. hyointestinalis LMG 2817	X67761	21
C. fetus LMG 6442	X67762	21
C. concisus LMG 7788	X67759	21

The PCR assay detected as few as one to five cells or chromosomes of C. jejuni SLV-428. The number of chromosomes was calculated according to the chromosomal size estimated by Karlyshev et al. (10). The detection of such a low number of cells is probably enhanced by the occurrence of three copies of the 23S rRNA genes on the chromosome (18).

Restriction enzymes specifically recognizing polymorphic areas within the amplified segment were identified with the WWWtacg version 2.38 software (Harry Mangalam, University of California, Irvine). Digestion with AluI gave unique combinations of fragments for C. jejuni and C. lari, respectively (Fig. 1). Isolates of C. upsaliensis showed either of two AluI banding patterns, one of which was the same as that for C. coli (Fig. 1). The two different AluI patterns of C. upsaliensis strains were expected, since both sequence variants have been published (Table 2). Differentiation between C. coli and C. upsaliensis was achieved by using a second enzyme, Tsp509I, which resulted in discriminatory banding patterns (Fig. 1). There was a discrepancy between the expected identity and the result of the PCR assay in two cases of 118 thermophilic campylobacters tested (Table 1). One human isolate previously identified as C. jejuni was typed as C. coli according to the restriction banding patterns. This isolate was positive for hippurate hydrolysis and was accordingly PCR positive for the hippuricase gene only present in this species (12). Thus, the PCR assay failed to classify this isolate correctly. The second isolate, a veterinary strain previously typed as C. coli, was according to the PCR assay C. jejuni. The PCR typing result was supported by a positive amplification of the hippuricase gene. However, the DNA fragment generated was about 300 bp shorter than expected (12). Nevertheless, all other C. coli isolates analyzed were negative for the hippuricase gene PCR, indicating that this particular isolate in fact is a C. jejuni strain with a truncated hippuricase gene, thus explaining the negative hippurate hydrolysis phenotype characteristic of strains of C. coli.

The PCR method presented in this study identifies with a single PCR low numbers of thermophilic campylobacters (1 to 5 CFU). In addition, by two simple restriction enzyme digestions performed directly in the PCR mixture, it differentiates between species within this group of bacteria. Thus, the method has the potential to be used for detection and identification of thermophilic campylobacters in complex samples, such as foods in which low numbers are present. This method also has the possibility of complementing or replacing phenotypic methods for identifying thermophilic Campylobacter species.