| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1Animal Sciences Group, Division of Infectious Diseases, and OIE Reference Laboratory for Campylobacteriosis, PO Box 65, 8200 AB, Lelystad, The Netherlands 2Keygene N.V., Department of Microbial Genomics, Wageningen, The Netherlands 3Department of Infectious Diseases and Immunology, Faculty of Veterinary Medicine, and OIE Reference Laboratory for Campylobacteriosis, Utrecht University, Utrecht, The Netherlands 4Pre-Harvest Food Safety and Enteric Diseases Research Unit, National Animal Disease Center, Agricultural Research Service, US Department of Agriculture, Ames, Iowa, USA
Correspondence Jaap A. Wagenaar jaap.wagenaar{at}wur.nl
Received 3 June 2005
Accepted 9 August 2005
Differentiation of Campylobacter fetus into C. fetus subsp. fetus (Cff) and C. fetus subsp. venerealis (Cfv) is important for both clinical and economic reasons. In the past, several molecular typing methods have been used for differentiation, including amplified fragment length polymorphism (AFLP). In this study, AFLP was employed to identify C. fetus subspecies specific markers that can serve as a basis for design of novel PCR primer sets for Cfv. Four groups of C. fetus strains with different phenotypic or genotypic traits were examined by AFLP using 22 different DdeI/MboI primer combinations. Specific AFLP fragments were deduced and sequenced resulting in 41 sequences. Based on the obtained sequences, five potential subspecies-specific PCR assays were developed. Extensive evaluation of the five selected PCRs with a set of 65 diverse C. fetus strains identified primer set Cf C05 as subspecies Cfv-specific. This newly developed PCR is fully consistent with the AFLP subspecies differentiation results. The data indicate AFLP as a powerful tool for comparing closely related genomes and for exploiting this information to develop a specific PCR with extensive typing potential.
Present address: PathoFinder B.V., Canisius Wilhelmina Ziekenhuis, Nijmegen, The Netherlands. Abbreviations: AFLP, amplified fragment length polymorphism; Cff, Campylobacter fetus subsp. fetus; Cfv, Campylobacter fetus subsp. venerealis.
The GenBank accession number for the target gene sequence of PCR Cf C05 is DQ146479.
 |
INTRODUCTION |
|---|
 TOP INTRODUCTION METHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
As some countries have eradicated Cfv from their cattle population, discrimination of both subspecies of C. fetus is crucial for veterinary control programmes. Subspecies differentiation is also of economic importance since Cfv-negative status of the animals is required by law for both the artificial insemination and embryo transfer industries (Anonymous, 2000; Wagenaar & Van Bergen, 2004).
The phenotypic test prescribed by the World Organisation for Animal Health (OIE) to differentiate the two C. fetus subspecies is the glycine-tolerance test (Wagenaar & Van Bergen, 2004). This test is based on the growth of Cff but not Cfv in 1 % glycine. However, the reproducibility of the assay is poor and the test can give ambiguous results (Harvey & Greenwood, 1983; On, 1996; On & Harrington, 2001). Other phenotypic tests, such as selenite reduction and antibiotic susceptibility, are only indicative and do not completely discriminate between the subspecies (On, 1996; Vandamme, 2000). Several molecular typing methods have been applied to discriminate the two subspecies, including PFGE (On & Harrington, 2001), amplified fragment length polymorphism (AFLP) (Wagenaar et al., 2001) and PCR (Hum et al., 1997; Wang et al., 2002). While numerical analysis of PFGE-DNA profiles is a valuable tool for differentiating C. fetus subspecies, it is time-consuming, difficult to interpret and therefore impractical for routine use. Similarly, the AFLP method is useful for differentiation of C. fetus subspecies (Wagenaar et al., 2001) but it is laborious, time-consuming and not available in every diagnostic laboratory. PCR is fast and easily applicable, but the currently available assays have been reported to give erroneous results (Hum et al., 1997; Wagenaar et al., 2001) or have only been tested for a very limited set of C. fetus strains (Wang et al., 2002). In the search for a robust, reliable and easy-to-use diagnostic PCR, we used the novel approach of DNA sequencing of C. fetus subspecies specific markers identified by AFLP to develop a PCR-based discriminative assay.
|
METHODS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
|
|
Chromosomal DNA of all strains was isolated with the Puregene kit (Gentra systems; BIOzym) and diluted to a final concentration of 40–50 ng µl–1 for each pool. For PCR on individual strains, the DNA of each strain was diluted to a concentration of 1–10 ng µl–1. PCR was also performed on cruder samples, i.e. on DNA prepared by 20 % Chelex-100 extraction (Engberg et al., 2000), and on cell suspensions (OD600 = 0.1 in SuperQ water).
Group selection and MboI/DdeI AFLP.
Based on results of AFLP genotyping and biochemical properties of the strains, four groups were identified (Table 2): group 1, Cff; group 2, Cff that differed from the first group in one HindIII/HhaI AFLP fragment; group 3, Cfv; and group 4, Cff strains that were biochemically aberrant compared to groups 1 and 2 [no growth in the glycine test indicating Cfv, whereas PCR (Hum et al., 1997) and AFLP (Wagenaar et al., 2001) indicated Cff]. From each identified group, two to four strains were selected from which DNA was pooled (indicated with pool number in Table 1). Each of the DNA pools was used as template in the AFLP analysis with restriction enzyme combination MboI/DdeI.
Identification of C. fetus subspecies specific markers.
Pooled DNA of the four C. fetus groups was examined by AFLP using the restriction enzyme combination MboI/DdeI. This enzyme combination was chosen based on the AFLPinSilico band pattern for the sequenced Campylobacter jejuni NCTC strain 11168 using the REcomb program (Reijans et al., 2003). AFLPinSilico mimics AFLP experiments and produces virtual fingerprints that enable the identification of fragments based on their length and the choice of selective nucleotides (Rombauts et al., 2003). Since the genome of C. fetus is not in the public domain, no information about suitable restriction enzymes for AFLP of C. fetus was available when this study was conducted. As C. jejuni is its nearest genome-sequenced relative, AFLPinSilico was done using this strain.
AFLP was performed essentially as described by Vos et al. (1995) and Van den Braak et al. (2004). Briefly, chromosomal DNA of each group was digested with restriction enzymes MboI and DdeI (Roche) and specific MboI and DdeI adaptors were ligated (MboI: forward 5'-CTCGTAGACTGCGTACC-3', reverse 5'-GATCGGTACGCAGTC TAC-3'; and DdeI: forward 5'-GACGATGAGTCCTGAG-3', reverse 5'-TNACTCAGGACTCAT-3'). Pre-amplification was performed as described by Vos et al. (1995), using primers without selective nucleotides. In the following selective amplification, samples were amplified using a 33P-labelled MboI primer containing one selective nucleotide, and a DdeI primer containing two selective nucleotides. In total, 22 different primer combinations were tested. Amplicons were run on 5 % denaturing polyacrylamide gels and fingerprints were visualized by phosphor imaging and screened automatically using AFLP QuantarPro software (Keygene Products). Fragments only commonly present in groups 1, 2 and 4 and absent in group 3 were potentially specific for Cff. Fragments only present in group 3 and absent in the other groups were potentially specific for Cfv. These candidate fragments were excised and re-amplified using AFLP primers without selective nucleotides. Successfully amplified products were sequenced on a capillary sequencer and analysed using the Basic Local Alignment Search Tool (BLASTX and BLASTN search) (http://www.ncbi.nlm.nih.gov/blast).
Development of subspecies-specific PCRs.
For all 41 properly sequenced fragments, primer sets were developed to enable specific amplification of the fragments. Each set of primers was examined for its subspecies specificity by pre-screening the set with DNA from a well-defined panel of seven Cff and seven Cfv strains, indicated in Table 1 by (Cff) and || (Cfv). Primer sets yielding a subspecies-specific fragment were used with DNA of all strains shown in Table 1. PCR was performed using the following conditions. Each 50 µl of reaction mixture contained 32.7 µl SuperQ water, 5 µl 10x PCR buffer II (Applied Biosystems), 3 µl 25 mM MgCl2 (Applied Biosystems), 0.3 µl 5 U µl–1 Taq polymerase (Applied Biosystems), 5 µl 2 mM dNTP (Amersham Biosciences), 1 µl 25 pmol µl–1 forward primer (Invitrogen, LifeTechnologies), 1 µl 25 pmol µl–1 reverse primer and 2 µl chromosomal DNA solution (1–10 ng µl–1). The PCR using primer set Cf C05 (forward, 5'-ATGATAAGATATATTTGTATCAG-3'; reverse, 5'-GAT GAAGAATATTACAAGATAAT-3') was performed as described above, except that 0.25 µl 5 U µl–1 SuperTaq polymerase (HT Biotechnology) and the supplied 10x buffer, including MgCl2, were used. All PCR reactions were performed in a Perkin Elmer 9700 thermocycler using the following conditions: 5 min incubation at 95 °C, 30 cycles of 1 min at 94 °C, 1 min at the calculated annealing temperature, and 2 min at 72 °C. Reactions were completed by extension at 72 °C for 10 min. The five selected PCRs (Cf B01, Cf B03, Cf B06, Cf C02 and Cf C05) with diagnostic potential were tested at both higher (+1 °C to +5 °C) and lower (–1 °C to –5 °C) annealing temperatures to assess the effect of temperature on the specificity of the PCR, and to reduce background. Higher annealing temperatures showing less background and identifying the same strains to be positive as the calculated annealing temperatures were selected. PCR Cf C05 had an optimum annealing temperature at 54 °C. All PCRs using the selected primer sets were performed in duplicate. PCRs that proved useful on chromosomal DNA were tested for their specificity on DNA extracted using the Chelex method and on cell suspensions. All PCR products exceeding 100 bp were run on a 1.5 % agarose gel and PCR products smaller than 100 bp were run on a 3 % agarose gel. PCR Cf C05 revealed a PCR product of 54 bp. PCR products were visualized by ethidium bromide staining.
|
RESULTS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
|
Development of subspecies-specific PCRs
To assess the use of the identified sequences for subspecies identification, PCR primer sets were designed for all 41 (5 Cff + 36 Cfv) sequences obtained and tested on a subset of C. fetus strains (Table 1, strains indicated by or ||). From the five selected Cff primer combinations, none were specific for Cff strains (Table 3). Yet, five of 41 primer combinations (Cf B01, Cf B03, Cf B06, Cf C02 and Cf C05) yielded reproducible PCR products for the specific identification of Cfv (Table 3). The diagnostic potential of these five Cfv-specific PCRs was further assessed with a set of 65 C. fetus strains, including three reference strains and 62 very diverse field isolates (Table 1). Strains originating from the same outbreaks displayed the same outcome for all individual PCRs (data not shown).
Comparison of the newly developed diagnostic PCRs with existing methods
To position the novel diagnostic PCR with existing AFLP and PCR assays, 65 strains were analysed with the different typing methods. In those assays, 24 out of 65 strains yielded a different classification in the various tests (Table 1). Reference strains ATCC 27374T (= NCTC 10842T), ATCC 19438T (= NCTC 10354T) and LMG 6570 (= CCUG 7477) gave similar results in all assays. With the other strains the results of the Cff-specific PCR described by Wang et al. (2002) gave the most variable results compared to all other tests. For this reason, this PCR method was not used further in the comparative study.
Further comparison of the other assays showed that the results of the newly developed primer sets Cf B01, Cf B03, Cf C02 and Cf B06 approached either those of the PCR described by Hum et al. (1997) or those of the AFLP. Primer sets Cf B01, Cf B03, Cf C02 and Cf B06 showed, respectively, two, four, three and nine discrepancies with the results of the PCR assay of Hum et al. (1997), and, respectively, eight, six, seven and one differences compared to the HindIII/HhaI AFLP. However, the results with the primer set Cf C05 were fully consistent (65/65) with AFLP, while eight of these strains yielded different results with the PCR of Hum et al. (1997).
As the PCRs with the primer set Cf C05 appeared to better reflect AFLP results, the use of this PCR was assessed with C. fetus target DNA extracted by Chelex and by cell suspensions rather than purified chromosomal DNA. All three DNA isolation methods yielded the same outcome for all individual strains resulting in acceptable PCR products (data not shown).
Sequence analysis of the subspecies-specific AFLP fragments
To gain more information about the nature of subspecies-specific genetic markers, the respective fragments were further analysed. All fragments amplified by primer sets Cf B01, Cf B03, Cf B06, Cf C02 and Cf C05 appeared to carry an open reading frame. BLAST analysis using publicly available databases, however, yielded no similarities with sequences present in the database. Furthermore, alignment of the DNA sequences of all five PCR products revealed no sequence similarities among the various sequences or with the Cfv-specific fragment amplified in the PCR by Hum et al. (1997). As the PCR Cf C05 was shown to be useful for differentiation, the target gene sequence was submitted to GenBank and given the accession number DQ146479.
|
DISCUSSION |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
At this time, only one assay is accepted by the OIE to discriminate between the two subspecies. Unfortunately, the reproducibility of this assay is poor. This is at least partially due to the sometimes slow and poor growth of C. fetus isolates. Whether this poor reproducibility reflects genetic differences is unclear and may await unravelling of the molecular basis underlying the glycine-tolerance test. An additional problem with the glycine-tolerance test is that this phenotype can be transferred between strains by phage-mediated transduction (Chang & Ogg, 1971). This makes the assays less attractive for subspecies differentiation purposes.
Reliable discrimination of the C. fetus subspecies is of veterinary importance as eradication programmes have been put in place in several countries. Due to the serious reproductive problems caused by Cfv, countries must report the presence of the causative agent of bovine genital campylobacteriosis (BGC) once diagnosed. Eradication programmes are fully dependent on the correct identification of the subspecies. More robust and easy-to-use discriminatory assays would facilitate the control of BGC. Several molecular tests have been evaluated as alternatives for the glycine test (Hum et al., 1997; On & Harrington, 2001; Wagenaar et al., 2001; Wang et al., 2002). Subspecies-specific PCRs have been described (Hum et al., 1997; Wang et al., 2002) but have been found to give erroneous results (Hum et al., 1997; Wagenaar et al., 2001). Development of new, more reliable, easy-to-perform tests is indispensable, and a multiplex PCR identifying a specific fragment for each subspecies would be a great asset. Due to the lack of the entire genome sequence of both subspecies, a genomic subtraction technique (Schober et al., 2001) which has been useful for identification of differences between strains (Winstanley, 2002) may be a valuable tool to detect genetic differences. In the present study, we used the power of the AFLP technique to identify differences among closely related genomes (Van den Braak et al., 2004) to successfully design a rapid PCR method for differentiation of C. fetus subspecies.
The first step towards the identification of subspecies-unique DNA fragments was the use of the appropriate restriction enzyme combinations in the AFLP. MboI/DdeI AFLP yielded 72 specific markers, including 59 for Cfv and 13 for Cff. The lower number of identified Cff-specific markers was probably due to the fact that the Cff-specific fragments must be shared by three different Cff groups, compared to one group for Cfv. AFLP differences may arise from large or minor genetic differences as well as from single point-mutations, which thus may cause false results. Although diagnostic PCR assays based on single nucleotide differences have been described (Misawa et al., 1998), larger sequence differences are likely to give more reliable results. For 41 of the 72 DNA fragments sequence information was obtained each of which had the potential to serve as a basis for development of subspecies-specific PCRs.
A rather low yield of subspecies-specific markers compared to the observed differences in MboI/DdeI AFLP was apparent. From the 41 PCR primer sets that were designed on the basis of the obtained specific sequences only five showed subspecies specificity towards the seven Cff and seven Cfv screening strains. This rather low yield may be explained by heterogeneity in the targeted sequence in different isolates. Furthermore, for each sequence only one primer set was developed. As a result of unsuitable primer sets, initially promising sequences might have been discarded, presumably reducing the number of final potential candidates. Another reason that these primer sets appeared nonspecific could be assigned to weak fragments, which were regarded negative in the MboI/DdeI AFLP, resulting in selection of false candidate markers. A further reduction from the five potential primer sets Cf B01, Cf B03, Cf B06, Cf C02 and Cf C05 to the one Cfv-specific primer set Cf C05 was obtained when evaluation of the primer sets used a much larger set of different strains than that screened in the initial selection of specific fragments.
A key part of our study was the evaluation of the usefulness of novel PCRs in the classification of a large number of field isolates. This analysis is complex, as the existing methods to distinguish Cfv and Cff, such as the glycine-tolerance test, are not optimal. Comparison of the results of AFLP, previously described PCR methods (Hum et al., 1997; Wang et al., 2002), and the newly developed PCRs yielded similar results when reference strains were used. With the diverse set of field strains, however, differences between the methods became apparent, indicating considerable strain diversity for the genetic markers used in the various assays. Yet, in our hands, the PCR with the primer combination Cf C05 yielded identical subspecies differentiation to the AFLP, which is considered a very discriminatory technique. Positive results were obtained with chromosomal DNA as well as with crude extracts, indicating the robustness of the assay. These data indicate that the Cf C05-based PCR is an important asset in C. fetus subspecies differentiation. For future diagnostics, a combination of an existing species-specific PCR and the Cf C05 PCR for subspecies discrimination may be optimal for C. fetus diagnostics.
The identification of subspecies-specific DNA fragments may provide important information about differences in virulence properties. We found no similarities for the obtained subspecies-specific sequences with proteins in the databases. Apparently, the sequences on which the developed PCRs were based have not previously been reported or identified in other bacteria. This could indicate that the new Cfv-specific PCR Cf C05 can be used for development of a diagnostic assay in which samples can be tested in a relatively high background of other bacteria.
New Cff-specific markers remained unidentified, and erroneous results were obtained with the PCR described by Wang et al. (2002). Therefore, attempts to characterize new Cff-specific markers by this or other molecular methods seem worthwhile. In this context it should be noted that only 22 out of the possible 64 (+1, +2) primer combinations have been used in the AFLP screening. In addition, other restriction enzyme combinations might yield a larger pool of putative specific markers to be tested.
Finally, it should be noted that our successful approach of using AFLP to identify specific markers adds a new dimension to this technique that may enhance the identification of future diagnostic targets as well as virulence genes.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  F. Schulze, A. Bagon, W. Muller, and H. Hotzel Identification of Campylobacter fetus Subspecies by Phenotypic Differentiation and PCR. J. Clin. Microbiol., June 1, 2006; 44(6): 2019 - 2024. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | J MED MICROBIOL | MICROBIOLOGY | J GEN VIROL | ALL SGM JOURNALS |
