OUP user menu

RAPD analysis of environmental, food and clinical isolates of Campylobacter spp.

Anthony C. Hilton, Debbie Mortiboy, Jeff G. Banks, Charles W. Penn
DOI: http://dx.doi.org/10.1111/j.1574-695X.1997.tb01036.x 119-124 First published online: 1 June 1997


The typing of Campylobacter is relatively poorly developed compared to that of the Enterobacteriaceae, and new molecular methods may provide useful approaches. The polymerase chain reaction was used to amplify randomly primed genomic DNA from Campylobacter isolates with an optimised randomly amplified polymorphic DNA protocol. Groups of isolates were analysed from chicken house environmental sources, chicken joints from retail sources, patients suffering from clinical disease and laboratory culture collections. Amplicons were separated by agarose gel electrophoresis, stained with ethidium bromide, and banding patterns captured in a digital form for computer analysis with GelCompar software. The method gave 100% typability and reproducibility for the isolates investigated and proved a useful technique for the epidemiological analysis of Campylobacter. Computer-based analysis of the randomly amplified polymorphic DNA generated profiles allowed relationships between isolates to be studied at the molecular level resulting in some indication of molecular correlates of the origins of isolates.

  • Campylobacter
  • Randomly amplified polymorphic DNA
  • Computer-aided epidemiological analysis
  • Molecular typing

1 Introduction

Campylobacter is recognised as the most significant cause of bacterial foodborne infection in developed countries [1]. The number of reported cases of Campylobacter enteritis has risen dramatically since 1977, Communicable Disease Surveillance Centre (CDSC) figures bringing the total laboratory reported UK cases in man to 44 414 during 1994, and 43 902 in 1995 [2]. The reported food poisoning cases are a crude estimation of the true extent of the burden of disease as in addition to difficulties arising from the isolation and characterisation of fastidious Campylobacter isolates, it is conservatively estimated that only 10% of cases of food poisoning are reported [3]. Associated costs run to hundreds of millions of pounds per year in terms of health care and loss of productivity [4].

The sources of foodborne outbreaks are rarely traced. In order to investigate sources and modes of transmission of these organisms it is important to have epidemiological typing methods which can discriminate between different strains but which can reliably and reproducibly recognise the same strain. Current methodologies for typing Campylobacter are poorly developed, and vary in their usefulness with respect to discriminatory power, cost and ease of use, with attention being turned to molecular methods in recent years.

A rapid typing method based upon randomly amplified polymorphic DNA (RAPD) was described [5, 6] and is a modification of the polymerase chain reaction (PCR) in which a single primer able to anneal and prime at multiple locations randomly distributed throughout the genome can produce a spectrum of amplified products characteristic of the template DNA. Thus, the entire genome is potentially accessible to priming and amplification. Detection of polymorphisms by established fingerprinting procedures requires prior knowledge of the target DNA sequence or cloned characterised probes. RAPD however can characterise anonymous genomes with arbitrarily chosen oligonucleotide primers [7]. Randomly amplified polymorphic DNA analysis has been reported for the characterisation of Campylobacter from various sources [813].

In this study a robust and reproducible RAPD method was developed and used as a molecular typing tool to distinguish strains of Campylobacter from various sources and computer aided analysis employed to investigate strain relatedness.

2 Materials and methods

2.1 Bacterial isolates

Campylobacter isolates from environmental (n=21), poultry (n=17), clinical (n=43) and culture collection (n=5) sources were purified where necessary by culture on selective media (LabM, Amersham, UK) by standard techniques. Isolates were confirmed as Campylobacter by Gram-stain, colonial and microscopic morphology, motility, catalase and oxidase tests. C. jejuni and C. coli strains were not differentiated by hippurate hydrolysis as previous work (unpublished) with the described RAPD primer showed profiles not to be species-specific. In addition, hydrolysis cannot always be relied upon due to the existence of hippurate negative C. jejuni strains [14] suggesting that this test does not correlate with overall genomic diversity. Isolates were grown in 5 ml brain heart infusion (BHI) broth (Unipath, UK) with shaking at 200 rpm for 24 h in a 5% O2:10% CO2:85% N2 atmosphere. Bacterial DNA template was prepared from these cultures as described by Mazurier et al. [12] with modifications. Briefly, 3 ml of culture was centrifuged in a benchtop microfuge (5 min at 13 000×g), the culture supernate discarded and the cells resuspended in 1 ml PBS to wash. The cells were centrifuged again and resuspended in 500 ml H2O, placed in a boiling water bath for 10 min, centrifuged again and the supernate removed to a fresh tube. The supernate A260 was read and diluted to give a DNA concentration of 10 ng ml−1.

2.2 RAPD reaction

The PCR was performed in a 25 µl volume containing 2.5 µl 10× PCR buffer (100 mM Tris-HCl, 35 mM MgCl2, 250 mM KCl, pH 8.8), 0.5 µl 10 mM dNTPs (Promega, Madison, WI, USA), 0.6 µl 100 mM Primer (1254: 5′-CCGCAGCCAA-3′, [15]), 1.25 units of Taq DNA polymerase (Gibco-BRL), 19.15 µl H2O and 2 µl DNA template at 10 ng µl−1. Reactions were overlaid with mineral oil and amplified in a PTC-100/60 thermal cycler (MJ Research, Watertown, MA, USA) as follows: one cycle of 4.5 min at 94°C followed by 5 low stringency cycles of 30 s at 94°C, 2 min at 21°C, 2 min at 72°C, and 35 high stringency cycles of 30 s at 94°C, 1 min at 32°C and 2 min at 72°C. The cycling was concluded with 5 min at 72°C and the reaction products stored at 4°C prior to analysis.

2.3 Profile analysis

Amplification products were detected by agarose gel electrophoresis with ethidium bromide staining and recorded by the IS500 digital imaging system (Flowgen, Kent). Digitally captured RAPD profiles were analysed with GelCompar software (Applied Maths, Belgium) with the band matching coefficient of Dice and UPGMA clustering to determine profile relatedness. In addition, the discriminatory ability of the described RAPD reaction was determined from Simpson's numerical index of diversity [16].

3 Results

An example of the electrophoretic profiles generated by RAPD analysis of Campylobacter is shown in Fig. 1. The profiles were analysed in duplicate to highlight the reproducibility of the described RAPD protocol. Data from longitudinal studies also suggest that the optimised RAPD reaction can reproducibly recognise the same strain analysed years later (data not shown). The RAPD analysis resulted in 100% typability, all isolates generating a specific profile of DNA fragments which was unique among the isolates investigated. Simpson's numerical index of diversity was calculated as 0.999. Profiles from all the isolates are represented in digital form in the dendrogram in Fig. 2. In addition, the relationships between isolates on the basis of RAPD fragments can be determined using the dendrogram. The analysis showed indications of clustering at 50% similarity level of the chicken meat isolates in three separate clusters, A, F and G, and a separate large cluster (B) incorporating a majority of the chicken house isolates. Clinical isolates comprised in some cases well-defined and separate clusters (E) while others were scattered among the other isolates examined. Two small heterogeneous clusters, C and D, contained a mixture of environmental and clinical isolates, and cluster G a mixture of clinical and poultry isolates. The culture collection strains were also dispersed among the identified clusters, the majority of Penner isolates being grouped in B while Penner 36 was in the heterogeneous group G, and NCTC 11828 in group A.

Figure 1

Example of RAPD profiles amplified from Campylobacter genomic DNA. Profiles were generated using the RAPD protocol detailed in Section 2. Analyses were run in duplicate for nine different strains. Lanes M are molecular weight markers (1 kb ladder, Gibco-BRL).

Figure 2

UPGMA dendrogram showing relationship between Campylobacter RAPD profiles. Profiles from all the isolates are represented in digital form. Clusters are indicated A–G. The horizontal scale indicates % similarity between profiles where vertical connections are made between strains or groups.

4 Discussion

In contrast to the uncertainties of characterisation based on variable phenotypic characteristics, the application of molecular typing methods can provide a stable and highly discriminatory analysis of bacterial isolates. Random amplification of polymorphic DNA has been successfully applied to the differentiation of thermophilic campylobacters and is shown here to provide 100% typability of isolates from clinical, food and environmental sources. DNase activity, which can give rise to untypability of isolates of Campylobacter by RAPD [10] was not observed in this study. The diversity of profiles, also observed in RAPD studies of Campylobacter from marine sources [10], represents the heterogeneity of the genus which probably arises in part from the natural competency of, and horizontal gene transfer between, campylobacters [17]. This is also reflected in the diversity index of 0.999. This indicates that if two strains were sampled randomly from the Campylobacter population, on 99.9% of occasions they would fall into different RAPD types. This is not only a measure of the high discriminatory power of the described RAPD method but a further indication of the heterogeneity of the Campylobacter genus. An index of >0.9 has been suggested as desirable if the results of a typing scheme are to be interpreted with confidence [16], and the described RAPD protocol fulfils this requirement. Since all of the Campylobacter isolates generated a unique RAPD profile there existed the possibility that the typing scheme was excessively discriminatory. This concern was addressed by analysing isolates which, by detailed serotyping, biotyping and epidemiological data, had been identified as the same strain. When characterised by RAPD the profiles generated from these isolates were identical (data not shown) confirming that the described RAPD protocol could reliably and reproducibly recognise the same strain.

Several monomorphic bands were also present in a majority of the isolates investigated. These fragments have been cloned and sequenced characterised (data not shown), however, in contrast to Salmonella isolates [18] the current lack of published genomic sequence data makes possible functional attributes of these genomic regions difficult to ascertain. This situation is likely to change in light of current proposals to sequence the Campylobacter genome which will allow valuable additional information to be gained from RAPD profiles.

Seven distinct clusters were observed at the 50% similarity level encompassing the isolates investigated. Clinical isolates are represented in each cluster which suggests that it is not a sub-population of campylobacters which are responsible for human disease, but a wide variety of strains. These data may also suggest that poultry isolates are not as diverse, being represented by two distinct clusters, however more isolates should be investigated.

A significant advantage to having genetic profiles of isolates on a computer database is that retrospective studies can be performed to allow epidemiological inferences about the sources and routes of transmission of Campylobacter. Currently this information is rarely available, however, RAPD profiles represent stable genetic markers that may provide useful indications of Campylobacter epidemiology and allow subsequent control measures to reduce the burden of disease and associated costs to society.

This method appears to be a valuable tool for the classification of Campylobacter isolates. Excellent typability and reproducibility coupled with relative technical ease make RAPD suitable for the epidemiological investigation of Campylobacter. In the long term, as sequence information becomes available, RAPD profiles may provide useful additional information and by profile analysis possible molecular correlates of the origins of isolates may be found.


This work was funded by a grant provided jointly between the Faculty of Science, University of Birmingham and Campden and Chorleywood Food Research Association.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
View Abstract