Bovine Campylobacter jejuni Strains Differ from Human and Chicken Strains in an Analysis of Certain Molecular Genetic Markers

Abstract

The association of four new genetic markers with a chicken, bovine, or human host was studied among 645 Campylobacter jejuni isolates. The γ-glutamate transpeptidase gene and dmsA were common in human and chicken isolates but uncommon among bovine isolates. In the t test, bovine isolates differed significantly (P < 0.05) from human and chicken isolates.

Campylobacter jejuni is a zoonotic human enteric pathogen with a large number of animal hosts (12, 19). Campylobacteriosis is a leading cause of human bacterial gastroenteritis in many industrialized countries (19). Epidemiological studies indicate that exposure to improperly cooked chicken meat, handling of raw chicken meat, and drinking unpasteurized milk are important risk factors for campylobacteriosis (12, 15, 19, 20).

The role of different animal sources in human infections is not well characterized. Molecular typing methods applied for fingerprinting of C. jejuni strains have shown overlapping genotypes between animal and human isolates (5, 16, 17, 21). Population biological studies using multilocus sequence typing (6) have revealed that a host-C. jejuni interaction may leave a signature in the bacterial genome. As a consequence, e.g., chicken- or cattle-associated populations can be assigned to their hosts (18). We investigated host association of C. jejuni isolates from cattle, chickens, and humans using PCR detection of four new genetic markers developed under our study. Using comparative genomics (3), four genetic markers—i.e., ggt, the γ-glutamyl transpeptidase gene; dmsA (Cju34), a subunit of the putative tripartite anaerobic dimethyl sulfoxide (DMSO) oxidoreductase (DMSO/trimethylamine N-oxide reductase) gene; Cj1585c, coding for a putative oxidoreductase; and CJJ81176-1371, a putative serine protease gene—were selected from the genomes of C. jejuni strains 81-176 (10), RM1221, and NCTC 11168. ggt is in the genome of 81-176 but not in the genome of NCTC 11168 or RM1221 (10). Gene Cj1585c of NCTC 11168 is replaced in 81-176 by a cluster of four genes (dmsA, dmsB, dmsC, and dmsD) (10). The presence of these four genes in a total of 645 C. jejuni isolates from bovine fecal samples (n = 131) (8), chicken cecal or meat samples (n = 205), and human patients (n = 309) (16, 17) was examined by PCR to find their suitability for host association studies. PCR primers designed for the amplification of the fragments are shown in Table 1. Twelve PCR products for each gene fragment were sequenced. The sequences of each gene were shown to be rather conserved (95.5 to 100% similarity within each gene) because only a few nucleotide positions (from 2 to 9) were found to be variable.

TABLE 1.

PCR primers used amplification of the fragments of the four marker genes

Gene (product)	Primer sequence		Product size (bp)
	Forward	Reverse
ggt (γ-glutamyl transpeptidase)	TTTTAGCCATATCCGCTGCT	AGCTGGAGTACCAGGAA	339
dmsAa	GATAGGGCATTGCGATGAGT	CTTGCTAGCCCAATCAGGAG	238
Cj1585c (oxidoreductase)	TGTTGTGGGTTTGCTGGATA	TTGCTTCACTGCATTCATCC	202
CJJ81176-1367/1371 (serine protease)	TGCAAAGCAGGGCTAAGAAT	TTATGGAGCTGGGGTGTTTC	318

^aCju34.

Statistical analyses were performed using SPSS software. The χ² test was used to test for similarity in the frequencies of marker genes within the isolates from different hosts. In addition, we used the paired two-tailed Student's t test for analysis of host associations for the combined set of four genes.

Frequencies of the genes are shown in Table 2. Similarly, the results of the paired two-tailed t test on the significance of the frequencies of the combined four genes from different hosts are shown in Table 2. These results indicated significant (P < 0.05) association of bovine and chicken isolates with their host source, but a high similarity was observed between the chicken and human isolates (P = 0.9949). Annual frequencies of the genes are presented for human isolates in Table 3 and for chicken isolates in Table 4. The analysis of the annual frequencies of the four genes combined showed that the human isolates were similar in 1996 and 2002 and 2002 and 2003, but differed between 1996 and 2003 (Table 3). The chicken isolates were similar in all study years (Table 4). These results revealed that these genes associated with metabolism and energy production (ggt, oxidoreductases) (2, 11, 22), colonization (ggt) (2, 11), or unknown function (serine protease genes) are not randomly distributed among the isolates from different hosts but show a host association.

TABLE 2.

Frequency of the four marker genes ggt, Cj1585c, dmsA (Cju34), and CJJ81176-1371 in 645 human, chicken, and cattle C. jejuni isolates

Marker gene (product)	No. of isolates with gene/total no. of isolates (%)			P value for sourcea:
	Human	Chicken	Bovine	Human/chicken	Chicken/bovine	Human/bovine
				χ2 test
ggt (γ-glutamyl transpeptidase)	169/309 (54.7)	75/205 (36.6)	11/131 (8.4)	<0.05	<0.05	<0.05
Cj1585c (oxidoreductase)	99/309 (32)	49/205 (23.9)	83/131 (62.6)	<0.05	<0.05	<0.05
dmsAb	256/309 (82.8)	151/205 (73.3)	18/131 (13.7)	<0.05	<0.05	<0.05
CJJ81176-1367/1371 (serine protease)	117/309 (37.8)	74/205 (36.1)	96/131 (73.3)	0.68	<0.05	<0.05
				t testc
				0.9949	0.0087	0.0122

^aP < 0.05 represents significant difference.

^bdmsA (Cju34) is a subunit of the putative tripartite anaerobic DMSO oxidoreductase gene.

^cSignificance (P < 0.05) of the frequency of the combined four genes by paired two-tailed t test.

TABLE 3.

Frequency of the four marker genes ggt, Cj1585c, dmsA (Cju34), and CJJ81176-1367/1371 in 309 C. jejuni isolates from humans

Marker gene (product)	No. of isolates with gene/total no. of isolates (%)			P value for yr:
	1996	2002	2003	1996-2002	1996-2003	2002-2003
				χ2 test
ggt (γ-glutamyl transpeptidase)	52/97 (53.6)	57/111 (51.3)	60/101 (59.4)	0.74	0.41	0.24
Cj1585c (oxidoreductase)	27/97 (27.8)	25/111 (22.5)	47/101 (46.5)	0.38	<0.05	<0.05
dmsAa	69/97 (71.3)	101/111 (91)	86/101 (85.1)	<0.05	<0.05	0.19
CJJ81176-1367/1371 (serine protease)	34/97 (35.1)	37/111 (33.3)	46/101 (45.5)	0.79	0.13	0.07
				t testb
				0.4506	0.0003	0.052

^admsA (Cju34) is a subunit of the putative tripartite anaerobic DMSO oxidoreductase gene.

^bSignificance (P < 0.05) of the frequency of the combined four genes.

TABLE 4.

Presence of the four marker genes ggt, Cj1585c, dmsA (Cju34), and CJJ81176-1367/1371 in 205 C. jejuni isolates from chickens

Marker gene (product)	No. of isolates with gene/total no. of isolates (%)			P value for yr:
	2003	2006	2007	2003-2006	2003-2007	2006-2007
				χ2 test
ggt (γ-glutamyl transpeptidase)	16/37 (43.2)	29/71 (40.8)	30/97 (30.9)	0.81	0.19	0.18
Cj1585c (oxidoreductase)	15/37 (40.5)	6/71 (8.5)	28/97 (28.9)	<0.05	0.21	<0.05
dmsAa	30/37 (81.1)	49/71 (69)	72/97 (74.2)	0.15	0.38	0.46
CJJ81176-1367/1371 (serine protease)	20/37 (54.1)	23/71 (32.4)	31/97 (31.9)	<0.05	<0.05	0.95
				t testb
				0.074	0.095	0.317

^admsA (Cju34) is a subunit of the putative tripartite anaerobic DMSO oxidoreductase gene.

^bSignificance (P < 0.05) of the frequency of the combined four genes.

The intestinal environments of cattle and chicken are quite different, which may select isolates with variable characteristics, e.g., related to energy metabolism, adaptation to lower or higher oxygen contents or amino acid metabolism. C. jejuni colonization in dairy cattle can be persistent, as shown by the studies in which the same genotype was isolated for up to 1 year (1, 13, 14). The life cycle of cattle is several years, providing a long potential time span for the adaptation of C. jejuni with its host. The life cycle of chickens, in contrast, is much shorter, 5 weeks or more. Our results suggested that host adaptation of certain C. jejuni strains is evident. The dmsA subunit was more often detected among chicken and human isolates than among bovine isolates (Table 2). In addition, dmsA-positive chicken isolates occurred with similar high annual frequency in 2003, 2006, and 2007 (Table 4), indicating that this characteristic is most probably important in colonization. The occasional significant annual fluctuation seen in the frequency of dmsA-positive human isolates may reflect variation in the infection sources (Table 3). In a recent study (9), dmsB was one of the genes present in C. jejuni strain A 74/C, shown to be robust colonizer in chickens, but absent from C. jejuni 11168(GS), a poorly colonizing strain (7). The C. jejuni NCTC 11168, 81116, and 81-176 strains have another putative DMSO oxidoreductase gene (homologous to Cj0264c) that differs from Cju34. In opposition, the Cj1585c-type oxidoreductase was more frequently present in isolates from cattle than in those from chickens or humans (Table 2). Analyses of C. jejuni genomes have predicted a branched complex electron transport chain capable of utilizing multiple electron donors and acceptors (22), and our results suggest flexibility in the oxidoreductase systems as well.

ggt (γ-glutamyl transpeptidase) has been shown to be important in the persistent colonization of C. jejuni in chickens (2), and recent studies (11) further extend the significance of this gene in the glutamine and glutathione metabolism and colonization of C. jejuni. In our study, the frequency of the ggt-positive human and chicken isolates was high (Table 2) and the frequencies remained similar over the study years (Tables 3 and 4). These results further reveal the importance of γ-glutamyl transpeptidase in colonization and pathogenesis. In contrast, a low frequency of ggt-positive isolates (8.4%) was found among bovine isolates (Table 2), suggesting that this type of metabolism is not crucial for colonization of the bovine gut. Similar variable frequencies to those in our study were found in the study by Barnes et al. (2).

The genomes of NCTC 11168, RM1221, and 81-176 have a subtilase-type serine protease gene homologous to CJJ81176-1367, which is located close to the CJJ81176-1371 gene in the genome of 81-176 (10). The G+C composition of this gene is 29%, whereas the G+C composition of CJJ81176-1371 is 36%, indicating that these genes most probably have different evolutionary origins. In our study, the serine gene was common among bovine isolates (Table 2) and less common among chicken and human isolates. The primers we used may amplify both types of the subtilase genes. Proteases in C. jejuni have a role in stress tolerance (4). Whether the serine protease is important in the pathogenesis of campylobacteriosis remains to be elucidated.

The genetic markers associated with metabolism, colonization, or an unknown protease function allowed assignment of the chicken or bovine source of C. jejuni. These results suggest that metabolic diversity is an important adaptive factor in host adaptation.