To partially address bias toward study of C. jejuni strains from high-income countries and the under-representation of strains from Asia in previous studies, we previously sequenced the genomes of 12 clinical isolates of C. jejuni from Asia: 4 from Thailand, 3 from Pakistan, and 5 from Vietnam (J. Harrison, unpub. data; Figure 1). We noted that 8 (67%) of these isolates possessed a cluster of genes homologous to the recently described T6SS (Figure 1). This finding was in contrast to findings regarding previously sequenced C. jejuni genomes; only 10 (14%) of 71 previously sequenced C. jejuni strains possessed an apparently intact T6SS gene cluster (Figure 1; full listing of genomes is in Technical Appendix Table 1). Several other strains from our study and previously sequenced strains contained >1 T6SS genes but not a complete T6SS cluster. Figure 1 shows the presence and absence of each T6SS gene in each available genome sequence (J. Harrison, unpub. data) and the previously sequenced strains. A nonrandom distribution of T6SS can be seen across the phylogenetic diversity of C. jejuni; T6SS is limited to certain clades, and degeneration of the T6SS gene cluster apparently occurs in parallel within several of those clades (Figure 1).

Our genome sequencing analysis indicated that strains possessing a complete T6SS cluster could be distinguished by the presence of the hcp gene (Figure 1) (9,10). Therefore, we used hcp as a proxy for determining the presence of a functional T6SS in 181 C. jejuni isolates from chickens, humans, and environmental sources (collections of the Oxford University Clinical Research Unit and the University of Exeter; online Technical Appendix Table 2). We designed and used a multiplex PCR (Technical Appendix Table 3) to screen for the presence of hcp in these isolates; the conserved C. jejuni housekeeping gene, gltA, was used as a positive control.

Of the 181 isolates, 28 originated from chickens in the United Kingdom and 21 from chickens in Vietnam. The hcp gene was found significantly more often in isolates from Vietnam (15 [71.4%] isolates) than in those from the United Kingdom (1 [3.5%] isolate) (p<0.01 by 2-sample Z-test; Technical Appendix Figure 1). An additional 38 of the isolates were from humans in the United Kingdom and 33 from humans in Vietnam; again, the hcp gene was significantly more prevalent in isolates from Vietnam (20 [60.6%] isolates) than those from the United Kingdom (1 [2.6%] isolate) (p<0.01 by 2-sample Z-test; Technical Appendix Figure 2).

We also found that patients infected with hcp-positive C. jejuni experienced bloody diarrhea more commonly than those infected with hcp-negative C. jejuni. For the 36 isolates for which detailed clinical data on patients were available, 6 (31.6%) of 19 patients in Vietnam who were infected with hcp-positive C. jejuni had bloody diarrhea, compared with 1 (5.9%) of 17 patients infected with hcp-negative C. jejuni (p<0.05 by 2-sample Z-test) (Figure 2). These results suggest a potential correlation between T6SS and bloody diarrhea, a serious clinical manifestation of the infection that results in higher rates of hospitalization and greater need for treatment with antimicrobial drugs (11). Moreover, Campylobacter-related septicemia developed in the 1 patient in the United Kingdom who was infected with a T6SS-positive strain (11). These data suggest that infection with the C. jejuni T6SS genotypic strains is associated with more severe disease. However, for sample bias to be ruled out, a comprehensive study is required in which the prevalence of T6SS is measured in C. jejuni samples from patients with mild and severe forms of infection.

We found a number of C. jejuni strains from humans and poultry that possessed the T6SS cluster, although some strains showed a slightly modified gene order (Technical Appendix Table 1 and Figure 3). However, most (61 [85.9%] of 71) of the previously sequenced C. jejuni isolates lacked a complete T6SS gene cluster (Figure 1); this finding might explain why T6SS was not discovered in C. jeuni sooner. Conversely, our PCR-based study frequently identified the hcp marker in isolates from Thailand, Pakistan, and Vietnam (Table). We cannot be certain that all of the isolates with the hcp marker possessed a complete and functional T6SS gene cluster, but the hcp gene is consistently associated with the presence of a complete T6SS cluster in all available sequenced C. jejuni genomes (Figure 1). This correlation lends confidence to the use of hcp as a proxy.

Poultry are a well-documented reservoir of human Campylobacter infection (12). We found that Campylobacter strains harboring the hcp marker were significantly associated with chickens in Asia. Large numbers of poultry are imported into North America and Europe from low-income countries, including Thailand (13). This process could introduce T6SS-positive Campylobacter genotypes into the food chains of the importing countries, posing a potential emerging threat to public health.

Our results suggest that the T6SS may be more prevalent in C. jejuni in Vietnam, Pakistan, and Thailand than in the United Kingdom. Furthermore, our results suggest that hcp may be a marker associated with severe human disease caused by C. jejuni infection in Vietnam, although there is no evidence that the association is causal. Chickens imported from these countries could be a source of hcp-positive strains and may have the potential to cause severe human infection.

Supplementary methods and primers used to identify hcp and gltA genes in Campylobacter jejuni, lists of C. jejuni strains included in analyses, prevalence of type-six secretion system (T6SS) genetic marker hcp in C. jejuni isolated from chickens and humans in Vietnam and the UK, and comparison of the gene orders in the T6SS gene clusters found in C. jejuni.