Application of high throughout sequencing to measure performance of commonly used selective cultivation methods for the food-borne pathogen campylobacter

Authors: Oakley, Brian; Morales, Cesar; Hiett, Kelli; Line, John; Seal, Bruce

Submitted to: FEMS Microbiology Ecology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 9/28/2011
Publication Date: 2/1/2012
Citation: Oakley, B., Morales, C., Hiett, K.L., Line, J.E., Seal, B.S. 2012. Application of high throughout sequencing to measure performance of commonly used selective cultivation methods for the food-borne pathogen campylobacter. FEMS Microbiology Ecology. 79(2):327-336.

Interpretive Summary: Campylobacter is an important food-borne human pathogen frequently associated with poultry products. The pathogen has traditionally been studied using a variety of bacterial cultivation methods that specifically select for the growth of this type of bacteria. Because there are many types of Campylobacter, some of which make people sick and others which may not, a long-standing problem has been how well different Campylobacter species grow with different culture methods. Only with the recent advances of next-generation, high-throughput DNA sequencing can we now fully answer this question. Such a sequencing approach, which basically provides a ‘bacterial census’, was used to determine how well commonly-used cultural methods work for diagnostic and survey purposes. The main question addressed was: 1. how well do different types of Campylobacter grow with each culture method? Five commonly used Campylobacter-selective plating media, two incubation atmospheres, and two incubation temperatures were compared to each other and to direct DNA sequencing of the poultry gut bacterial microbiome using 16S rRNA gene tagged-pyrosequencing. The growth media were generally very selective in that 88 to 97 percent of DNA sequences from the media were classified as Campylobacter despite the fact that only a small proportion (less than 0.04 percent) of the DNA sequences in the original sample used for plating was Campylobacter. One of the methods which is based on filtration instead of having antibiotics in the media formed fewer Campylobacter colonies and had the highest proportion of non-Campylobacter sequences, but the relative proportions of Campylobacter sequence types were very similar for all conditions tested. The five media used here may essentially be equivalent for use to culture campylobacters. Importantly, more contaminating bacteria were found when culturing was completed at 42 versus 37 degrees C. These contaminants were relatively rare, but it may be better to culture for campylobacters at 37 degrees C rather than 42 for this sample type. Finally, a number of new Campylobacter types were found using the newer DNA sequencing technology, suggesting that there may be many more Campylobacter species than are currently recognized by traditional culturing techniques. This information will be useful for diagnostics and regulatory agencies to help determine how to detect campylobacters in environmental samples.

Technical Abstract: Campylobacter is an important food-borne human pathogen which has traditionally been studied using a variety of selective cultivation methods. Here we use next-generation sequencing to ask: 1) How selective are commonly-used Campylobacter cultivation methods relative to the initial community being screened, and 2) How do the specificity and sensitivity of these methods compare to one another. We compared bacterial recovery in a factorial design with five commonly used Campylobacter-selective plating media (Cefex, Capetown, MCCDA, CLA, and CVA), two incubation atmospheres (10 percent CO2, 5 percent O2, 85 percent N2 and 10 percent CO2, 10 percent H2, 80 percent N2), and two incubation temperatures (37 and 42 degrees C), and compared these results to direct sequencing of the poultry gut microbiome using 16S rRNA tagged-pyrosequencing. Analysis of 404,104 raw sequence reads revealed that the fecal bacterial community was dominated by Bacteroides, Brevibacterium, and Lactobacillus, while Campylobacter encountered in the feces represented only a small proportion (less than 0.04 percent) of sequences. The growth media were highly selective in that 88 to 97 percent of media DNA sequences were most closely related to Campylobacter spp. The Capetown method which relies on filtration instead of antibiotics recovered significantly fewer CFU (p<2e-16) and had the highest proportion of non-Campylobacter sequences (12%), but recovery of Campylobacter spp. sequence types was very similar across all five media. Of the nine most abundant Campylobacter OTUs, which accounted for at least 0.01 percent of sequences, all nine were recovered on each medium, in similar proportions. Incubation atmosphere had little effect on recovery, but a significant difference in media specificity (more non-Campylobacter OTUs; p equals 0.028) was found for media incubations at 42 versus 37 C. These results provide insight into the bacterial community structure of the poultry gut microbiome and suggest that commonly used media selective for Campylobacter spp. generally perform equivalently but are biased for certain types of Campylobacters, some of which remain unknown. Detection of several novel 16S rRNA OTUs and MLST sequence types most closely related to C. jejuni support a previously unrecognized diversity of cultivable Campylobacters.