|Home | About | Journals | Submit | Contact Us | Français|
Botulinum neurotoxin (BoNT), the most toxic substance known, is produced by the spore-forming bacterium Clostridium botulinum and, in rare cases, also by some strains of Clostridium butyricum and Clostridium baratii. The standard procedure for definitive detection of BoNT-producing clostridia is a culture method combined with neurotoxin detection using a standard mouse bioassay (SMB). The SMB is highly sensitive and specific, but it is expensive and time-consuming and there are ethical concerns due to use of laboratory animals. PCR provides a rapid alternative for initial screening for BoNT-producing clostridia. In this study, a previously described multiplex PCR assay was modified to detect all type A, B, E, and F neurotoxin genes in isolated strains and in clinical, food, environmental samples. This assay includes an internal amplification control. The effectiveness of the multiplex PCR method for detecting clostridia possessing type A, B, E, and F neurotoxin genes was evaluated by direct comparison with the SMB. This method showed 100% inclusivity and 100% exclusivity when 182 BoNT-producing clostridia and 21 other bacterial strains were used. The relative accuracy of the multiplex PCR and SMB was evaluated using 532 clinical, food, and environmental samples and was estimated to be 99.2%. The multiplex PCR was also used to investigate 110 freshly collected food and environmental samples, and 4 of the 110 samples (3.6%) were positive for BoNT-encoding genes.
Botulinum neurotoxins (BoNTs) are the most toxic agents known, and as little as 30 ng neurotoxin is potentially lethal to humans (36). These toxins are responsible for botulism, a disease characterized by flaccid paralysis. Seven antigenically distinct BoNTs are known (types A to G), and BoNT types A, B, E, and F are the principal types associated with human botulism (37). Significant sequence diversity and antigenically variable subtypes have recently been reported for the type A, B, and E neurotoxin genes (14, 22, 23, 42).
Apart from the species Clostridium botulinum, which itself consists of four phylogenetically distinct groups of organisms, some strains of other clostridia, namely Clostridium butyricum and Clostridium baratii, are also known to produce BoNTs (2, 4, 7, 13, 20, 26, 34, 44). Also, strains that produce two toxins and strains carrying silent toxin genes have been reported (8, 22, 24, 39). Due to the great physiological variation of the BoNT-producing clostridia, their isolation and identification cannot depend solely on biochemical characteristics (32). Indeed, the standard culture methods take into consideration only C. botulinum and not C. baratii and C. butyricum, and identification and confirmation require detection of BoNT by a standard mouse bioassay (SMB) (12). The SMB is highly sensitive and specific but also expensive, time-consuming, and undesirable because of the use of experimental animals. Detection of neurotoxin gene fragments by PCR is a rapid alternative method for detection and typing of BoNT-producing clostridia (3). Different PCR methods have been described for detecting neurotoxin type A-, B-, E-, and F-producing clostridia (9, 15-18, 21, 40, 41).
A previously described multiplex PCR method able to simultaneously detect type A, B, E, and F neurotoxin genes is a useful tool for rapid detection of the BoNT-producing clostridia (31). While this method generally has a high level of inclusivity for detection of type B, E, and F neurotoxin genes, limitations for detection of the recently described subtype A2, A3, and A4 strains have been identified (6, 28). To increase the efficiency of this multiplex PCR method, new primers were designed to detect genes for all identified type A neurotoxin subtypes (19). Additionally, an internal amplification control (IAC) was added according to ISO 22174/2005. The specificity and selectivity of this multiplex PCR method were evaluated in comparison with an SMB (12) using target and nontarget strains, and the robustness was assessed using clinical, food, and environmental samples. Moreover, to evaluate the applicability of this multiplex PCR method, a survey with food and environmental samples was performed in a German food control laboratory.
The inclusivity and exclusivity of the primer s (Table (Table1)1) were tested using a number of bacterial strains (Table (Table2).2). All BoNT-producing clostridia were previously typed using an SMB (12). Clostridium strains were cultured in 9 ml of Trypticase peptone glucose yeast extract (TPGY) broth and incubated under anaerobic conditions using jars and an anaerobic gas-generating kit (Oxoid, Basingstoke, United Kingdom) at 30°C for 24 h. Campylobacter jejuni was cultured in Bolton broth (Oxoid) and incubated under microaerophilic conditions at 41.5°C for 48 h. All other strains were cultured in brain heart infusion broth (Oxoid) and incubated under aerobic conditions at 37°C for 24 h.
Aliquots (1 ml) of enrichment broth were subjected to DNA extraction using Chelex 100 matrix (Bio-Rad, Hercules, CA) (19) and used as templates for multiplex PCR.
A total of 532 samples (150 stool, 234 honey, 143 food, and 5 environmental samples) collected from botulism outbreaks and environmental surveys in the past 10 years in Italy were used to define the robustness of the multiplex PCR method. The food samples included 41 samples of canned vegetables, 89 samples of shellfish, 7 samples of canned meat, 2 samples of sausages in oil, 2 samples of canned tuna fish, and 2 samples of anchovies in brine. The presence of BoNT-producing clostridia was determined using a culture method described below combined with (i) an SMB (12) and (ii) the multiplex PCR method. All samples collected before 2004 were stored at −20°C, defrosted at the time of this study, analyzed by the multiplex PCR, and reanalyzed by an SMB. Samples collected after 2004 were analyzed on the day of arrival in the laboratory by both the SMB and the multiplex PCR method.
One gram of each sample was cultured in 9 ml of TPGY broth and incubated under anaerobic conditions at 30°C. After 24 h, 1-ml aliquots of the enrichment broth were subjected to DNA extraction using Chelex 100 matrix (19) and used as templates for the multiplex PCR. The broth was reincubated in jars at 30°C. After an additional 72-h enrichment period, the broth was subjected to an SMB to detect BoNTs. The broth media that were PCR negative at 24 h were tested again with the multiplex PCR at 96 h.
An additional 110 food and environmental samples collected in Germany were analyzed using the multiplex PCR method after enrichment in TPGY broth incubated under anaerobic conditions at 30°C for 4 days.
The SMB was performed in accordance with European Directive 86/609/EEC on the protection of animals used for experimental and other scientific purposes.
In order to identify false-negative results, an IAC was constructed using the procedure of Abdulmawjood et al. (1) and the commercial plasmid pUC19 (GenBank accession number L09137) as a template. Briefly, the primers used to amplify the IAC, designed using Primer Express 1.5 software (Applied Biosystems, Foster City, CA), generate a 698-bp product (Fig. (Fig.1).1). The IAC was built in competition with the primers amplifying the type F neurotoxin gene (29). One microliter of pUC19 was transferred to a vial containing 50 μl of a mixture of 50 mM KCl, 2.5 mM MgCl2, 10 mM Tris-HCl (pH 8.3), 200 μM of each deoxynucleoside triphosphate, 1 μM primer IACf and 1 μM primer IACr (Table (Table1),1), and 2.5 U of Taq polymerase (Applied Biosystems, Roche Molecular Systems). A 30-cycle PCR was carried out using denaturation at 95°C for 1 min, annealing at 55°C for 1 min, and elongation at 72°C for 2 min, followed by a final extension at 72°C for 10 min.
Optimization of the multiplex PCR protocol was performed in two laboratories. At Istituto Superiore di Sanità (Italy), it was performed with C. botulinum type A strain ATCC 19397, C. botulinum type B strain ATCC 27765, C. butyricum type E strain ATCC 43755, and C. botulinum type F strain NCTC 10281 using a My-Cycler thermal cycler (Bio-Rad). At the Bavarian Health and Food Safety Authority (Germany), optimization was performed with C. botulinum type A strain NCTC 7272, C. botulinum type B strain REB 1944 (Strain Collection of the Consultant Laboratory for Anaerobic Bacteria, University of Leipzig, Leipzig, Germany), C. butyricum type E strain REB 1718, and C. botulinum type F strain REB 1955 using a Mastercycler Gradient thermal cycler (Eppendorf, Hamburg, Germany). In both cases, the PCR was performed with a 50-μl mixture containing 2× multiplex PCR master mixture (Qiagen, Hilden, Germany), 0.3 μM of each primer (Table (Table1),1), and 3 μl of purified DNA template from each reference strain. The reaction mixture was heated at 95°C for 15 min to activate the hot-start Taq polymerase and then subjected to 30 cycles of denaturation at 95°C for 30 s, annealing at temperatures ramped from 50 to 60°C for 30 s, and extension at 72°C for 30 s, followed by a final extension at 72°C for 7 min. Further tests were carried out to optimize the annealing and extension times and the required numbers of amplification cycles. Annealing and extension times of 30, 60, and 90 s were each tested in combination with 30, 35, and 40 cycles of amplification. Optimization of the copy number of the IAC included in each PCR was performed as reported elsewhere (19).
A 50-μl reaction mixture contained 2× multiplex PCR master mixture, 0.3 μM of each primer (Table (Table1),1), 3 μl of purified DNA template, and 1,500 copies of IAC. The reaction mixture was heated at 95°C for 15 min to activate the Taq polymerase and then subjected to 35 cycles of denaturation at 95°C for 30 s, annealing at 56°C for 30 s, and extension at 72°C for 90 s, followed by a final extension at 72°C for 7 min.
PCR products were analyzed by 2% agarose gel electrophoresis at 90 V for 70 min, and the sizes of the fragments were determined by relating their positions on the gel to those of standard DNA fragments (EZ Load 100-bp molecular ruler; Bio-Rad Laboratories, Hercules, CA) (Fig. (Fig.1).1). To avoid contamination, sample preparation, DNA amplification, and electrophoresis were carried out in three different rooms.
The inclusivity, exclusivity, and relative accuracy of the multiplex PCR were calculated using the Microval protocol (5). Inclusivity is the ability of the PCR method to detect the target analyte in a wide range of strains. Exclusivity is the lack of interference with the PCR methods by a relevant range of nontarget strains. The relative accuracy is the degree of correspondence between the response obtained by the cultural method and the results of the multiplex PCR with identical samples (33). Statistical analysis of the results was performed with the McNemar χ2 test (43). A value equal to or greater than 3.84 indicates significance at the 0.05 level.
The amplification conditions that gave the best resolution of all five amplicons included annealing at 56°C for 30 s and 35 cycles of amplification when the My-Cycler was used and annealing at 55°C for 30 s and 36 cycles of amplification when the Mastercycler Gradient was used (data not shown).
Using the optimized PCR, it was determined that 1,500 copies of the IAC was the lowest number that gave reproducible IAC amplification. Using this number of copies of the IAC, the amplicon-specific band was present in the type A, B, and E neurotoxin gene-positive cultures and absent in the type F neurotoxin gene-positive cultures (Fig. (Fig.11).
With 182 target and 21 nontarget strains, 100% inclusivity and 100% exclusivity were found for the multiplex PCR when it was compared to the SMB (Table (Table22).
Table Table33 shows the results of a comparison of the multiplex PCR and the SMB using 532 naturally contaminated samples collected in Italy. A total of 170 samples were positive in the SMB, and 168 samples were positive in the PCR analysis. One hundred twenty-eight samples were positive in the PCR analysis after 24 h of enrichment, while a further 40 samples were positive in the PCR analysis after 4 days of enrichment. The relative accuracy of the multiplex PCR was 99.2%, and only four stool samples showed discordant results with the SMB (three samples were positive with the SMB and negative with the PCR, and one sample was negative with the SMB and positive with the PCR). McNemar's test demonstrated that there are no statistically significant differences (χ2 = 0.25; P > 0.05) between the results obtained for stool, food, and environmental samples by the cultural method and the results obtained by the multiplex PCR. An additional three samples negative in the SMB were not included in this comparison because the IAC was not amplified, indicating that there was PCR inhibition (Table (Table33).
The multiplex PCR was then used to ascertain the presence of BoNT-producing clostridia in 110 food and environmental samples from Germany. Four samples (3.6%) were positive (three honey samples for the type B neurotoxin gene and one soil sample for the type E neurotoxin gene), and the IAC was amplified in all samples.
The multiplex PCR has the advantage of simultaneous detection of several clostridia possessing type A, B, E, and/or F botulinum toxin gene(s). The multiplex PCR described by Lindström et al. (31) was used as the starting point for the present work. Since this multiplex PCR was described, a considerable number of neurotoxin subtypes have been described (35). While the multiplex PCR described by Lindström et al. provided good detection of type B, E, and F botulinum toxin genes, it did not always provide good detection of all type A botulinum toxin genes. The nucleotide sequences of primers CBMLA1 and CBMLA2 (31) were aligned with sequences of all type A neurotoxin subtypes (GenBank accession numbers for subtype A1, M30196, X52066, AF461540, and AF488749; GenBank accession numbers for subtype A2, X73423, AY953275, and DQ310546; GenBank accession number for subtype A3, DQ185900; GenBank accession number for subtype A4, DQ185901; GenBank accession number for subtype A5, EU679004). While the primers are conserved for the A1 and recently isolated A5 subtypes (11), there were mismatches with sequences for subtypes A2, A3, and A4. Therefore, we used another previously described type A primer set (19) in the multiplex PCR assay and detected all subtypes of type A neurotoxin genes despite the fact that primer IOA 1 had two mismatches with the sequence for subtype A5 (GenBank accession number EU679004) and three mismatches with the sequence for subtype A4 (GenBank accession number EU341307).
Optimization of the annealing temperature and the number of cycles is an important step in the development of a multiplex PCR assay (25, 27). The inclusion of the new type A primer pair, originally intended for a real-time PCR (20), in the multiplex PCR made it necessary to make modifications to the original multiplex PCR conditions described by Lindström et al. (31). These modifications include a decreased primer annealing temperature (56°C instead of 60°C) and an increased number of cycles (35 cycles instead of 27 cycles).
A high level of PCR inhibition was found in a study of the prevalence of C. botulinum in food raw materials used in refrigerated processed foods with extended durability (10). To avoid false-negative results in the present study, an IAC competitive with the type F neurotoxin gene was synthesized and included in the PCR mixture. Use of 1,500 copies of the IAC did not interfere with detection of the type F neurotoxin gene.
Inhibitory substances were present in only three stool samples in which the IAC was not amplified. All these samples were negative in the SMB. These stool samples were from botulism patients treated with active carbon. Efficacy of the sample preparation steps used to extract DNA from bacteria and to eliminate the presence of inhibitors is crucial in the development of a standardized test (30, 38). An in-house Chelex 100 DNA extraction method was selected for this study as it was previously shown to be faster, to be simpler to perform, and to cost less than commercial extraction kits (19).
The relative accuracy of the multiplex PCR method (99.2%) was higher than that reported for two previously described PCR-based surveys. One survey of the presence of C. botulinum in food samples produced a relative accuracy of 95.6% (16), while in a second survey of fish and environmental samples the relative accuracy was 88.9% (18).
The multiplex PCR for detection of BoNT-producing clostridia has high levels of specificity (100%) and relative accuracy (99.2%). Moreover, this method is faster and considerably less expensive than an SMB. The effectiveness of this method has been confirmed using a large number of strains and clinical, food, and environmental samples.
The results of this study show that the multiplex PCR method can be used to test clinical, food, and environmental samples in outbreak situations or used in routine surveillance studies to determine the presence of neurotoxin-producing clostridia. This method is not a substitute for the SMB, but it can be used to reduce the number of animal tests and to deliver results more rapidly and cheaply than an SMB in an initial screening step. When used together, the multiplex PCR and SMB can be used to distinguish neurotoxin genes that do and do not form active neurotoxin. A major advantage of the multiplex PCR method, like the original multiplex PCR described by Lindström et al. (31), is its ability to detect and readily differentiate BoNT genes encoding the four neurotoxins (types A, B, E, and F) associated with human botulism.
This work was supported in part by the Italy-USA project “Infant Botulism,” by the Finnish Ministry of Agriculture and Forestry, and by a Competitive Strategic Grant from the BBSRC.
Published ahead of print on 14 August 2009.