The real-time PCR asays developed in this study provide an effective, rapide, and sensitive alternative method to culture-based methods for the detection and the quantification of C. coli and C. jejuni in pure cultures and in complex samples.
To use real-time PCR for quantitative measurements and to ensure a correct quantification, information on both linear range and amplification efficiency of the real-time PCR assay must be available. With a quantitative detection limit of 10 genome copies, an amplification efficiency of 99%, and a linear range of seven to eight orders of magnitude, the C. coli and C. jejuni real-time PCR assays allowed a precise quantification of C. coli or C. jejuni DNA amounts extracted from pure culture preparations. The specificity of the assays was assessed (i) by the species-specific amplification of DNA from different field strains/isolates of C. coli and C. jejuni, and (ii) by the absence of amplification from DNA isolated from 30 pig faecal, feed, and environmental samples previously determined to be Campylobacter-free by culture. The real-time PCR assays were also shown to be highly specific since no PCR amplicons were detected when the method was applied to DNA from different bacterial reference strains, including different Campylobacter species, Campylobacter-related bacteria, and other bacteria. Both intra- and inter-assay coefficients of variation of the Ct values for the purified genomic DNA were satisfactorily low and in concordance with those reported for other molecular assays based on PCR amplification . They confirmed the reliability and the accuracy of the technical setup over time and over the complete range of quantification.
The technique was developed to detect and quantify C. coli and/or C. jejuni directly in pig faecal, feed, and environmental samples. In order to determine the detection limits of C. coli and C. jejuni real-time PCR assays for field samples, Campylobacter-negative faecal samples were spiked with 10-fold dilutions of the Campylobacter suspensions of each reference strain (C. jejuni NCTC 11168 and C. coli CIP 70.81). Standard curves for environmental and feed samples were constructed in a similar way. The established C. coli and C. jejuni real-time PCR assays proved highly sensitive (with a quantitative detection limit of approximately 2.5 × 102 CFU/g of faeces, 1.3 × 102 CFU/g of feed, and 1.0 × 103 CFU/m2 for the environmental samples) and were linear over a range of six orders of magnitude (from 2.0 × 102 to 2.0 × 107 CFU/g of faeces). Both intra- and inter-assay coefficients of variation of the Ct values for the DNA extracted from Campylobacter-negative faecal samples did not differ significantly. This may indicate that the main reason for variation is not due to pipetting errors in setting up the PCR assay but may be caused by contaminants from the fecal samples. Nevertheless, we did not observe systematically lower CV values of intra- and inter-assay variations with purified genomic DNA. This does not support the hypothesis that inhibitors and contaminants may interfere with uniform and consistent dilution as well as the amplification of target DNA.
Samples tested in this study constitute complex biological substrates due to the presence of (i) numerous types of bacteria, (ii) different kinds of inhibitors, and (iii) food degradation products [36, 37]. Moreover, contrary to faecal and caecal chicken samples [35, 38], the consistency and the composition of pig faecal samples are highly variable and heterogeneous (i) between individuals, (ii) over time according to the age of the animals, and (iii) depending on the diet components in the same way as for cattle faeces [39, 40]. In this study, we sampled faeces of sows, piglets, weaners, and finishers, exhibiting considerable heterogeneity (water content, presence of mucus, and fiber content). All these variables may have an impact on the DNA extraction process and inhibitor removal, affecting the quality and the quantity of DNA obtained, thereby limiting the sensitivity of molecular studies. The modified sample preparation procedure, which included (i) a large volume of faeces (5 g fresh weight), (ii) a boiling step known to remove inhibitors of the Taq polymerase , and (iii) the use of a DNA extraction kit, allowed a better homogenization of the faeces and achieved partial removal of inhibitors. No difference was noticed between real-time PCR assays and culture at both qualitative and quantitative levels for faecal samples differing by the composition, the consistency, or the age of the sampled animal (data not shown). Nevertheless, in this study, the potential presence of PCR inhibitory compounds was in parallel assessed with the use of an internal bacterial control of extraction and amplification in a separate real-time PCR test . Inhibitors of real-time PCR were identified only in 4% of the examined samples, which were consequently removed from the quantification study. Moreover, the DNA extraction step reproducibility, an important parameter when evaluating the DNA purification , was satisfactory proved with the low CV values of the inter-assay variability including the DNA extraction procedure.
Three faecal samples of experimentally infected pigs, detected as negative by PCR and direct streaking, were positive by culture after an enrichment step (one out of 41 and two out of 26 for C. coli and C. jejuni real-time PCR assays respectively) leading to a sensitivity of 97.6% and 92.3%. Although the internal control was positive, we cannot exclude the hypothesis of inhibition of C. coli and C. jejuni amplification. Indeed, it was previously reported that some PCR primers are more markedly affected than others by impurities present in DNA preparations [43, 44]. Moreover, it could be false negative PCR samples, which have been below the detection limit of the two real-time PCR assays. Genetic variability among the isolates of Campylobacter spp., which has been demonstrated previously [45–47], can also affect the efficacy of a PCR if changes occur within one or both of the binding sites [17, 46]. The enrichment step enhanced the sensitivity of the bacteriological method by lowering the detection limit. Nevertheless, even if it is helpful for poorly contaminated samples, researchers have reported several cases in which C. jejuni signals detected by direct PCR disappeared after enrichment. Conversely C. coli signals were maintained when present before enrichment, or else became detectable when undetectable before enrichment [24, 48]. This suggests that the enrichment media may favour the growth of one Campylobacter species comparatively to the other .
Furthermore, for the experimentally infected pigs, only one culture-negative faecal sample was positive by real-time PCR for each target leading to a specificity of 96.2% for both C. coli and C. jejuni real-time PCR assays. These results may be due to the presence of viable but nonculturable (VBNC) forms or dead bacteria cells, since DNA-based tests detect all DNA of the extract from live as well as dead bacteria [27, 29, 50]. If this is the case, it is another advantage of these real-time PCR assays as Campylobacter cells in a VBNC state may potentially be still infectious [18, 51]. The bacteriological method may also explain these results given that the sensitivity of culture may vary depending on the Campylobacter spp. due to differences in susceptibility to antibiotics present in selective agar . Moreover, in pig faceal and environmental samples, the enrichment of C. jejuni could be difficult due to the presence of a high background flora and due to the more numerous C. coli quantity .
Finally, for the faecal samples of experimentally infected pigs, we observed a good correlation at the quantitative level between culture enumeration and quantitative PCR for both C. coli and C. jejuni real-time PCR assays (R2 = 0.90 and R2 = 0.93 respectively). Among the PCR-culture positive samples, the real-time PCR quantification seems to be accurate compared to the culture enumeration used as a gold standard. Indeed, more than 95% of the samples with a difference in cell number of less than 2 logs, of these 72.5% and 67% less than 1 log respectively for C. coli and C. jejuni real-time PCR assays. The observed discrepancy might be due to the possible presence of VBNC forms, dead cells and antagonistic bacterial species. Another possibility could be the impact of dilution factors used for quantitative culture or an insufficient homogenization of the samples. This method provides a mean to identify and quantify at the species level C. coli and C. jejuni directly from faecal, feed, and environmental samples without requiring an enrichment step. For the different field samples tested, the qualitative data (specificity and sensitivity) as well as the quantification results obtained by C. coli real-time PCR matched equally the results obtained by bacterial culture. In this study, no C. jejuni was identified among the faecal, feed, and environmental samples from the different pig herds by conventional PCR or by our C. jejuni real-time PCR assay. Conversely, all the Campylobacter tested were identified as C. coli by both methods. In France, pigs were found to be almost always contaminated by C. coli, these first results confirmed this predominance. Nevertheless, given that we can find both species in pigs [10, 12–14], these real-time PCR assays allow a direct and rapid investigation of the carriage and the excretion of C. coli and C. jejuni in conventional pigs.