Evaluation of two PCR-based procedures for typing Clostridium perfringens

INTRODUCTION Clostridium perfringens, a widely distributed microorganism found in soil and the alimentary tract of animals, is associated with a wide variety of diseases affecting most domestic animal species and humans (Table 1). The microorganism is classified into 5 types (A, B, C, D and E) according to the production of 4 major lethal toxins (Table 2). Types A, B, C and D occur most commonly in domestic animals. Confirmation of the diagnosis of a disease caused by C. perfringens relies on the assessment of clinical signs and post mortem findings in combination with the isolation, determination of pre-formed toxins in intestinal contents and identification and typing of the organism. The organism can be visualised microscopically on smears of affected tissues and intestinal contents as a Gram-positive stout rod. Bacteriological identification is conducted by anaerobic culturing of affected tissues and intestinal contents on blood agar or on selective agar media. The toxins are identified by means of the toxin-neutralisation test in the skin of guinea-pigs or neutralisation tests in mice. These conventional methods are timeconsuming, expensive and require the use of live animals. Apart from ethical considerations, the guinea-pig intradermal test (GPI) lacks precision, is technically difficult to perform and also requires experience to analyse and interpret. Since the elucidation of genes encoding for the 4 major toxins of C. perfringens, many PCR-based detection methods have been developed and used for genotyping of C. perfringens. Two approaches are commonly used: PCR based on individual reactions for each toxin gene and multiplex PCR for the simultaneous amplification of the major toxin genes in 1 tube. The 1st method, although time-consuming and costly, seems to be more suited to detecting the organism directly from clinical or environmental samples, whereas the multiplex procedure produces better results on cultured material . To assess whether the PCR-based genotyping of C. perfringens could be a better substitute to conventional in vivo methods used in our laboratory, a multiplex PCR assay (MP) and a PCR based on the individual detection of each toxin gene (IP) were evaluated and compared to results obtained with the conventional methods on isolates and overnight cultures from clinical specimens.


INTRODUCTION
Clostridium perfringens, a widely distributed microorganism found in soil and the alimentary tract of animals, is associated with a wide variety of diseases affecting most domestic animal species and humans (Table 1).The microorganism is classified into 5 types (A, B, C, D and E) according to the production of 4 major lethal toxins (Table 2).Types A, B, C and D occur most commonly in domestic animals 4,6 .
Confirmation of the diagnosis of a disease caused by C. perfringens relies on the assessment of clinical signs and post mortem findings in combination with the isolation, determination of pre-formed toxins in intestinal contents and identification and typing of the organism 1,4,6 .The organism can be visualised microscopically on smears of affected tissues and intestinal contents as a Gram-positive stout rod.Bacteriological identification is conducted by anaerobic culturing of affected tissues and intestinal contents on blood agar or on selective agar media.The toxins are identified by means of the toxin-neutralisation test in the skin of guinea-pigs or neutralisation tests in mice 7 .
These conventional methods are timeconsuming, expensive and require the use of live animals.Apart from ethical considerations, the guinea-pig intradermal test (GPI) lacks precision, is technically difficult to perform and also requires experience to analyse and interpret.
Since the elucidation of genes encoding for the 4 major toxins of C. perfringens, many PCR-based detection methods have been developed and used for genotyping of C. perfringens 2,3,8 .Two approaches are commonly used: PCR based on individual reactions for each toxin gene and multiplex PCR for the simultaneous amplification of the major toxin genes in 1 tube 3 .The 1st method, although time-consuming and costly, seems to be more suited to detecting the organism directly from clinical or environmental samples, whereas the multiplex procedure produces better results on cultured material 2,3,8 .
To assess whether the PCR-based genotyping of C. perfringens could be a better substitute to conventional in vivo methods used in our laboratory, a multiplex PCR assay (MP) and a PCR based on the individual detection of each toxin gene (IP) were evaluated and compared to results obtained with the conventional 5 ' -ATAAT C C C AAT C AT C C C AAC T ATG-3'), the beta-toxin gene (5'-AGGAG GTTTTTTTATGAAG-3' and 5'-TCTAAA TAGCTGTTACTTTGT-3') and epsilon-toxin gene (5'-TACTCATACTGTG GGAACTTCGATACAAGC-3' and 5'-CT CAT C T C C CATAAC T G CAC TATAAT TTCC-3') were selected for the IP protocol.Ten µ of the templates from isolates were used with each set of primers (each to a final concentration of 0.8 µM), 250 µM of each dNTP, 10 mM Tris-HCl pH 8.8, 1.5 mM MgCl2, 50 mM KCl, 0.1 % Triton X-100 and 1 unit of Taq polymerase (Takara Biomedicals) in a 50 µ reaction volume.
Amplification was conducted under the following conditions: 2 min at 93 °C followed by 30 cycles of 30 sec at 93 °C, 30 sec at 48 °C and 30 sec at 72 °C for mixes with primers specific for the alpha-and epsilon-toxin genes.Reaction mixes containing primers for the beta gene were denatured for 2 min at 93 °C, followed by 30 cycles of 45 sec at 93 °C, 45 sec at 50 °C and 45 sec at 72 °C.
All amplicons were visualised under UV illumination after being subjected to electrophoresis for 20 min in a 1.5 % agarose gel (Roche Molecular Biochemicals) with 0.5 µg/m of ethidium bromide.
Template DNA was extracted from C. botulinum, C. septicum, C.tetani, C. sordelli and C. chauvoei using the same protocol as for C. perfringens.The extracted DNA was used in PCR under the same conditions as for C. perfringens specimens to determine the specificity.
The GPI test was carried out according to the standard protocol 7 .

RESULTS
Both PCR protocols yielded the expected amplification products when used on C. perfringens isolates (Figs 1, 2).For the MP, the products were of the following sizes: 324 bp for the alpha-toxin gene, 196 bp for the beta-toxin gene and 655 bp for the epsilon-toxin gene.For the IP assay, the product sizes were 247 bp, 1025 bp and 403 bp for the alpha-, betaand epsilon-toxin genes, respectively.All reference isolates used in the evaluation yielded the expected amplicons.
Products of similar sizes to the reference isolates were obtained when the 2 PCR assays were used on cultures from clinical samples.The PCR results obtained were similar to those described earlier 3,8 and confirmed the good sensitivity and specificity of the PCR assays.PCR assays appear to be a suitable alternative typing method for C. perfringens, especially with regard to clinical samples.
Cultures from clinical samples were subjected to PCR after visual identification of C. perfringens colonies.They were also subjected to the GPI test.MP, IP and GPI tests results are summarised in Table 3. Except for 2 specimens with negative results, all samples yielded amplicons that could be characterised as corresponding to the alpha-, beta-or epsilontoxin genes, allowing the isolates to be typed as types A, B, C or D.
No product resulted from DNA templates prepared from the other clostridial species, confirming the specificity of the PCR assays.

DISCUSSION
Except for the 3 clinical samples with equivocal results on the GPI test, there was similarity in results of the 43 samples typed by the GPI-, the IP-and the MP-based procedures.Samples 41 and 45 gave weak positive results in the GPI test, but specific for type D. With the GPI test, 5 of 46 clinical samples (10.8 %) either showed equivocal results or showed weak skin reactions in guinea-pigs used for typing.The 5 samples yielded clear results with the PCR assays.In vivo typing methods for C. perfringens are known to some degree to lack precision and usually require an experienced diagnostician.This condition has limited the diagnostic capability of many laboratories, so that the PCR seems a better alternative typing method.Clinical specimens used in this evaluation were cultured before PCR amplification.No attempt was made to directly amplify DNA extracted from field specimens.C. perfringens forms part of the normal flora of the intestinal tract of animals and man.It can invade the parenchymatous organs shortly after death.C. perfringens is only considered to be significant if it is present in large numbers in the intestine, and if its isolation is correlated with the clinical signs and lesions seen at necropsy 5 .For bacterial evaluation, a heavy growth of C. perfringens is defined as a virtually pure growth obtained from both the initial inoculum and at least the 1st quadrant when streaked out using the normal method 1 .

Sample MP-based typing IP-based typing GPI-based typing
Performing the PCR assay after bacterial growth has been evaluated, therefore, becomes more appropriate in a diagnostic context.The use of multiplex PCR rather than single gene assays (more suited for direct detection of the organism in field samples) provides a better alternative to in vivo toxin typing.The combination of PCR with culture for identification of the organism allows a qualitative and quantitative evaluation of the specimen, and a more rapid and accurate typing of the C. perfringens type involved.

Table 3 :
Comparison of Clostridium perfringens typing results using the multiplex PCR (MP), the individual gene amplification PCRs (IP) and the guinea-pig intradermal test (GPI) on cultures from clinical specimens.

Table 2 : The major toxins of Clostridium perfringens.
unequivocally typed by PCR.Samples 39 and 40 were negative on both tests.They originated from cases clinically suspected to be associated with C. perfringens infection, but never confirmed.