Clinical samples from 71 bovine from different Brazilian states were processed for the analysis of anaerobe organisms with emphasis on the isolation and characterization of Clostridium spp. From these, eighty-nine Clostridium perfringens strains were recovered: 32 from liver, 19 from intestinal contents, 14 from kidney, 6 from rumen, 5 from nervous system, 4 from bone marrow, 2 from udder tract, blood, spleen and lung, and one from muscle. Four reference Clostridium perfringens types A, B, C, and D were used as controls in this study. All isolates were cultivated in appropriate media, and after centrifugation the supernatant and sediment were separated. From pure supernatant post exopolysaccharide (EPS) extraction, mouse toxigenicity tests were performed, determining protein and protein plus carbohydrate, respectively. ELISA was performed from sediments. The results showed that 51 (57.3%) of the isolates were toxigenic to mice when inoculated by intraperitoneal route; bacteria from different organs had variable patterns of toxigenicity. Toxigenicity of EPS extracts was only expressed when protein concentration was 0.04 mg/mL and between 0.31 and 0.5 mg/mL for carbohydrate. Isolates were characterized as toxigenic when showing optimum protein and carbohydrate concentrations.
Clostridium perfringens; toxigenicity; bovine isolates; ELISA
TOXIGENICITY CHARACTERIZATION OF Clostridium perfringens FROM BOVINE ISOLATES
L. BALDASSI1 CORRESPONDENCE TO: L. Baldassi Laboratório de Bacteriologia Geral e Micobacterioses, Instituto Biológico. Avenida Conselheiro Rodrigues Alves, 1252, 04014-002, São Paulo, SP, Brasil. E-mail: firstname.lastname@example.org , M. L. BARBOSA2, E. E. BACH3, S. T. IARIA4
1 Laboratório de Bacteriologia Geral e Micobacterioses, Instituto Biológico; 2 Instituto Adolfo Lutz; 3 Departamento de Química UNICASTELO/UNESP; 4 Departamento de Microbiologia, Instituto de Ciências Biomedicina, USP.
ABSTRACT: Clinical samples from 71 bovine from different Brazilian states were processed for the analysis of anaerobe organisms with emphasis on the isolation and characterization of Clostridium spp. From these, eighty-nine Clostridium perfringens strains were recovered: 32 from liver, 19 from intestinal contents, 14 from kidney, 6 from rumen, 5 from nervous system, 4 from bone marrow, 2 from udder tract, blood, spleen and lung, and one from muscle. Four reference Clostridium perfringens types A, B, C, and D were used as controls in this study. All isolates were cultivated in appropriate media, and after centrifugation the supernatant and sediment were separated. From pure supernatant post exopolysaccharide (EPS) extraction, mouse toxigenicity tests were performed, determining protein and protein plus carbohydrate, respectively. ELISA was performed from sediments. The results showed that 51 (57.3%) of the isolates were toxigenic to mice when inoculated by intraperitoneal route; bacteria from different organs had variable patterns of toxigenicity. Toxigenicity of EPS extracts was only expressed when protein concentration was 0.04 mg/mL and between 0.31 and 0.5 mg/mL for carbohydrate. Isolates were characterized as toxigenic when showing optimum protein and carbohydrate concentrations.
KEY WORDS:Clostridium perfringens, toxigenicity, bovine isolates, ELISA.
Clostridium perfringens belongs to the Clostridium genus that includes many gram-positive anaerobe bacteria, a group that is sensitive to oxygen and has the ability to form heat-resistant endospores.
This organism is characterized by the ability to produce numerous extracellular toxins and is classified into five types (A to E). Bacteria are notoriously prolific toxin producers; there are at least 17 different reported bacteria (31,32) associated with pathogenicity (21,25,26). Clostridium perfringens is responsible for many different extra- and intraintestinal diseases, such as gas gangrene, necrotic enteritis, animal enterotoxemia, and human food poisoning (21,25,29).
Although enterotoxemia diagnosis is usually based on history, clinical signs, and necropsy findings, laboratory analysis is essential to confirm the presence of toxins and to properly identify the pathogen (31,34).
Characterization of C. perfringens and its toxins is well established, although few data are available in Brazilian literature about its prevalence related to human and animal diseases.
Several laboratory methods have been used to characterize C. perfringens, such as enzyme linked immunosorbent assay (ELISA), bacteriocin production, and toxigenicity for Vero cells, but none is sufficiently adequate to identify all C. perfringens strains (27).
This study was performed to investigate isolates by using different tests in order to obtain a better method to substitute the in vivo test.
MATERIALS AND METHODS
Samples were collected from 71 bovine with clinical manifestation of from many Brazilian states between 1996 and 1998.
These samples were initially cultivated in Cooked Meat Broth (CMB) and incubated at 37°C for 48 hours. From these cultures, 0.1mL loop aliquot was streaked on 5% sheep blood agar and incubated in anaerobic conditions at 37°C for 24 hours. Colonies showing related characteristics of C. perfringens form, aspect, color, and hemolysis were submitted to Gram stain; colonies corresponding to Gram-positive bacilli were cultivated in CMB in the same above conditions. After the incubation period, cultures were submitted to additional tests including catalase, lecithinase and gelatinase production, and glucose, lactose, and skimmed milk fermentation used to species identification. Strains identified as C. perfringens were then sub cultivated in CMB and stored at 4°C.
We used 89 isolates of C. perfringens (32 recovered from liver, 19 from intestinal contents, 14 from kidney, 6 from rumen, 5 from nervous system, 4 from bone marrow, 2 from udder tract, blood, spleen, lung, and one from muscle) were tested. Four reference strains from American Type Culture Collection (ATCC) numbers 3624 (type A), 3626 (type B), 3628 (type C), and 3629 (type D) were used as controls.
The 89 isolates were used in toxigenicity determination, imunoenzymatic assay (ELISA), and exopolysaccharide (EPS) analysis.
Protein extraction and toxigenicity test
After incubation under anaerobic conditions on plates with 5% sheep blood agar, 5 colonies of each plate were transferred to Triptose Yeast Extract Broth (TYB) and further incubated in anaerobic conditions. The whole volume (10mL) of TYB was transferred to a fresh TYB (90mL) and incubated according to Pons et al. (24). The purity of each culture was analyzed by Gram stain method. Cultures were then centrifuged at 7.500 xg for 15 minutes at 4°C, and supernatants were submitted to protein - Bradford (4), adapted for microtechnique quantification (30,33). A standard curve was based on concentrations of 20, 40, 60, 80, 100, and 120 mg/mL bovine serum albumin (BSA). The OD value at 500nm was read in a microplate reader (BIO-RAD, model 3550-UV) coupled to a computer using Microplate Manager Program Protein extracts were tested for toxigenicity.
Toxigenicity was determined by the IP inoculation of each protein extract supernatant into 10 mice, as follows: five animals with 0.5mL of pure extract and the other five with 0.5 mL of extract treated with 1% trypsin (final concentration) after incubation at 37°C for 30 minutes.
Exopolysaccharide extraction and toxigenicity determination
Fifty milliliters from each TYB culture were freeze-dried using a lyophilizator (CHRISS, model Delta 1A), diluted 1:4 with 95% ethanol, and maintained at 4°C for 24 hours. Sediments obtained after centrifugation (3.000xg for 10 minutes) and total ethanol evaporation were re-suspended in sterilized distilled water. Evaluation of carbohydrate concentration was performed according to Dische (11) and read by spectrophotometer (Pye Unicam) through 620 nm-filter. The standard curve was prepared with a glucose solution at concentrations of 63, 125, and 250 mg/mL. Protein concentration was determined using these extracts.
Toxigenicity was determined by the IP inoculation of each EPS extract into 5 mice. All inoculated animals were observed over 96 hours for clinical manifestations or death (3).
Enzyme-linked immunosorbent assay (ELISA)
Antibody was obtained from New Zealand rabbits (protein extraction culture of each strain). Rabbits were immunized with 0.5 mL of antigen-sediment re-suspended in 0.5mL of 0.85% saline solution and homogenized with 0.5mL of Freund complete adjuvant in lymphonods (1). The animals were bled after 30 days.
Sediments were washed twice with tris-glycin buffer, 60mM, pH 8.7 (24), and suspended in 1mL of the same buffer, triturated on liquid nitrogen, and stored at -20°C until ELISA was performed.
Each triturated and washed sediment was diluted 1:10 in carbonate buffer solution, pH 9.6, distributed in four wells of polystyrene microplates, and incubated at 4ºC for 24 hours as per Bach et al. (1). The plates were washed three times with 200 mL of phosphate buffer solution (PBS), pH 7.4, with 0.05% of Tween 20. Then, 200 mL of diluted antibody (1:5 in PBS) of each type (A, B, C, and D) was added to each corresponding well, and the plates incubated at 37ºC for 4 hours. The plates were washed again and re-incubated with 200 mL of sheep anti-rabbit IgG conjugate marked with phosphatase (Sigma). After another wash, the wells received 200 mL of p-nitrophenyl phosphate buffer and were incubated at 37°C for 30 minutes. The reaction was interrupted by the addition of 200 mL of NaOH 3M, and absorbency was measured at A405 nm using ELISA reader (BIO-RAD, model 3550-UV) coupled with a computer with Microplate Manager Program.
For test optimization, antigens were analyzed at different concentrations (600, 300, and 150mg BSA/mL) and conjugated in 1:5 and 1:10 dilutions. These results were replicated three times and compared with negative controls (reagent and normal serum) (28,37).
From the 89 C. perfringens isolates, 14 (15.7%) were not toxic when treated with trypsin, 8 (9.0%) were toxic only after trypsinization, 29 (32.6%) with both treatments, and 38 (42.7%) were not toxic.
EPS extracts inoculated into mice showed 12 (13.5%) deaths, 9 with treated or non-treated treated EPS extracts, and 3 with non-treated extracts. Quantification of carbohydrate extracts of these strains varied from 0.31 to 4.74 mg/mL, and protein concentration from 0.04 to 0.36 mg/mL.
These results indicated that the extracts with 0.04 mg/mL protein killed mice and were then considered toxigenic. In the same samples, carbohydrate concentration varied from 0.31 to 0.59 mg/mL. The relationship carbohydrate/protein varied from 7.75 to 14.75 mg/mL. This ratio, however, does not seem to be important because values ranging from 3.33 to 64.40 mg/mL were not responsible for death in those animals (Table 1).
Carbohydrate (mg/mL) and protein (mg/mL) quantification and mouse toxicity of 89 C. perfringens EPS extracts.
At first, ELISA was performed using as antigens (Ag) C. perfringens types A, B, C, and D, reference strains with 600 mg of proteins and goat anti-rabbit immunoglobulin G conjugated with alkaline phosphatase at a dilution 1:5. Antisera (Ab) were produced in rabbits against C. perfringens isolates (ATCC) types A, B, C, and D. ELISA demonstrated low specificity and crossreactivity against all antigens. Antiserum type D (Ab) recognized its own antigen, showing crossreactivity with the other antigens. No reaction was observed with normal serum nor with the reagents employed (Table 2).
ELISA OD values for all extracts using antisera against ATCC isolates types A, B, C, and D.
For the 89 isolates, ELISA was also performed the in the same optimization. Table 3 shows that one isolate (1.1%) was classified as type C, 36 (40.4%) as A, 30 (33.7%) as D, and 9 (10.1%) as B. Thirteen isolates (14.6%) could not be classified.
Classification of 89 C. perfringens isolates into different types by ELISA OD values.
Diagnosis of enterotoxemia in animals depends on historical data, clinical signs, and necropsy findings. Laboratory confirmation is essential for agent determination and toxin detection (38). Characteristics such as gelatinase production, nitrate reduction, motility absence, lactose and stormy skimmed milk fermentation can identify C. perfringens species (19,20,38), while toxin determination can indicate the circulating types (6,35).
The e toxin and enterotoxin have their toxic activity increased when treated with trypsin (13,23,39), while the b is inactivated by the same enzyme (12). These data are in agreement with our study, where 9% of the isolates showed toxigenicity after trypsin treatment, and 15.7% lost toxigenicity with the same treatment, although toxigenic isolates were found in all organs.
Of the 89 strains tested, 57.3% were toxin producers, therefore, potentially pathogenic; this is in agreement with Tsai et al. (34), who reported 60% of bovine C. perfringens isolated from slaughterhouses in Canada as toxigenic.
These observations are in agreement with Holliday (15), who associated the degree of encapsulation with pathogenicity, explained by the observation of higher toxigenic strains from clinical fecal isolates or from atmospheric origin. Bacteria can lose their capsules or become wrinkled (16) when maintained for a long time or submitted to many passages in artificial media. These observations justify the occurrence of a toxigenic strain in one organ and atoxigenic in another in the same animal. In relation to enterotoxin, Mahony et al. (20) demonstrated that the strain ability to produce toxin in vitro does not reflect its capacity to produce it in vivo. Therefore, toxin detection in laboratory cultures, maintained under the same conditions, can determine the toxigenic potential of a certain strain.
The scheme for serological typing of C. perfringens was initially developed in England, using 8 isolates from type A (14). Hughes et al. (16) using the same scheme same with 57 antisera obtained 65% of typing in food poisoning brotes and 59% in cases of gas gangrene and other infections caused by this agent. Mahony et al. (19,20), Klotz (18), and Willis (39) reported that the agglutination test demonstrated that there were non-typified isolates in four types. However, according to Watson (38), in epidemiological surveys it would be necessary to use 139 different antisera (75 English and 64 Japanese strains), not available commercially and still to have conditions for preparing it for other strains non-typible with this group of antisera. The International Serotyping Scheme describes 82% of success in the classification of C. perfringens strains with 143 antisera (17).
The existence of atoxigenic variants or those that have lost their toxigenicity besides the fact that the serological test related to the toxins are not clearly differentiated by serological tests for C. perfringens (5,9).
To perform this study, we used antisera produced against bacterial cells of the four ATCC pattern types of C. perfringens (20), as the use of non-commercially available antitoxins results in many difficulties in their production and purification; this led to trying a new ELISA technique to identify serological types of C. perfringens.
Table 2 shows that ELISA demonstrated crossreactivity among samples that can be seen due to the presence of antigens with group-specific polysaccharides that are lost when isolates become wrinkled (8). The high level of crossreactivity can justify the use of test strains isolated at different times and artificially maintained.
According to Calvert et al. (7) and Dazzo (10 ), exopolysaccharides (EPS) are antigenic molecules and can be related to pathogenicity (22,36). As per Bach and Guzzo (2), EPS are found in cultures of phytopathogenic bacteria and are responsible for antigenicity. In this study, EPS seems to be influencing pathogenicity since the strains that caused death, when the inoculum was from protein extraction, were the same ones that caused death in EPS.
EPS extraction method was shown to be sensitive enough to detect considerable levels of carbohydrates in supernatants of C. perfringens strain cultures, although small protein levels were still present.
Toxigenicity revealed by animal death should be related much more to protein rather than to carbohydrate levels. The relationship carbohydrate/protein does not seem to be important because in the samples that did not cause death, these values varied between zero and 64.40 mg/mL. The values 7.75 to 14.75 mg/mL that determined death were within this range. The values found for carbohydrate, 0.66 - 4.75 mg/mL are also inside of the parameters that caused death (0.31- 0.59 mg/mL). In relation to protein, the values found in EPS extraction were between 0.05 and 0.36 mg/mL for the strains whose extracts did not determine mouse death, and 0.04 mg/mL for those that caused death of the inoculated animals.
Received March 5, 2001
Accepted June 21, 2001
1BACH EE., ALBA APC., RODRIGUES NETO J. Detection of strain for Xanthomonas campestris pv. Citri (Hasse) Dye by enzyme-linked immunosorbent assay (ELISA). Fitopatol. Bras, 1982, 7, 407-15.
2BACH EE., GUZZO SD. Serological relationship and chemical composition of exopolysaccharides (EPS) from different pathotype of Xanthomonas campestris pv. citri and X. campestris pv. manihotis. J. Phytopathol.,1994, 141, 1-9.
3BALDASSI L. Isolamento de bactérias do gênero Clostridium e detecção de toxina botulínica a partir de materiais obtidos de bovinos com suspeita clínica de botulismo. São Paulo: Universidade de São Paulo, Faculdade de Saúde Pública, 1986. 59p. [Dissertação Mestrado].
4BRADFORD MM. A rapid and sensitive method for the quantification of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem., 1976, 72, 248-54.
5BUDDLE MB. Degraded strains of Clostridium welchii type C isolated from sheep in New Zealand. J. Comp. Pathol. Ther., 1954, 64, 223-4.
6BULLEN JJ. Role of toxins in host-parasite relationships. In: AJL SJ., KADIS S., MONTIE TC. Microbial toxins. New York: Academic Press, 1970:.233-76. v.1
7CALVERT HE., LALONDE M., BHUVANESWARI TV., BAUER WD. Role of lectins in plant-microorganism interactions. 4. Ultrastructural localization of soybean lectin binding sites on Rhizobium japonicum. Can. J. Microbiol.,1978, 24, 785-93.
8CHERNIAK R., HENDERSON BG. Immunochemistry of the capsular polysaccharides from Clostridium perfringens: selected Hobbs strains 1, 5, 9 and 10. Infect. Immun., 1972, 6, 32.
9DALLING T., ROSS HE. Clostridium welchii: notes on the relationship between the types of cultures and the production of toxin. J. Comp. Pathol. Ther., 1938, 51, 235-49.
10DAZZO FB. Adsorption of microorganisms to roots and other plant surfaces. In: BITTON G., MARSHALL KC. Adsorption of microorganisms to surfaces. New York:.Wiley, 1979: 253-316.
11DISCHE Z. General color reactions. In: WHISTLER RL., WOLFROM ML. Methods in carbohydrate chemistry. New York: Academic Press, 1965: 478-81.
12DOWELL VR., HAWKINS TM. Laboratory methods in anaerobic bacteriology. CDC Laboratory manual. D.H.E.W. Publ, 1974 (74-8272).
13GRANUM PE., WHITAKER JR., SKJELKVALE R. Trypsin activation of enterotoxin from spores of Clostridium perfringens type A. Biochim. Biophys. acta, 1981, 668, 325-32.
14HOBBS BC., OAKLEY CL., WARRACK GH., CRUICKSHANK JC. Clostridium welchii food poisoning. J. Hyg., 1953, 51, 75-101.
15HOLLIDAY MG. Rapid identification of Clostridium perfringens by counter-immunoelectrophoresis. Med. Lab. Sci., 1985, 42, 322-5.
16HUGHES JA., TURNBULL PCB., STRINGER MF. A serotyping system for Clostridium welchii (C. perfringens) type A, and studies on the type-specific antigens. J. Med. Microbiol., 1976, 9, 475-84.
17INTERNATIONAL COMMISSION ON MICROBIOLOGICAL SPECIFICATIONS FOR FOOD OF THE INTERNATIONAL UNION OF BIOLOGICAL SOCIETIES FOR FOODS (ICMSP). Clostridium perfringens. In: JAMES & JAMES micro-organisms in foods 5. Microbiological specifications of food pathogens. London: Chapman & Hall, 1996: 112-25.
18KLOTZ AW. Application of FA techniques to detection of Clostridium perfringens. Public Health Rep., 1965, 80, 305-11.
19MAHONY DE., SWANTEE CA. Bacteriocin typing of Clostridium perfringens in human feces. J. Clin. Microbiol, 1978, 7 , 307-9.
20MAHONY DE., AHMED R., JACKSON SG. Multiple typing techniques applied to a Clostridium perfringens food poisoning outbreak. J. Appl. Bacteriol., 1992, 72, 309-14.
21MCDONEL JL. Toxins of Clostridium perfringens types A, B, C, D In: DORNER F., DREWS J. Eds. Pharmacology of bacterial toxins. New York: Pergamon, 1986: 477-517.
22NIEPOLD F., HUBER SJ. Surface antigens of Pseudomonas syringae pv. syringae are associated with pathogenicity. Physiol. Mol. Plant Pathol., 1988, 33, 459-71.
23OAKLEY CL., WARRACK GH. Routine typing of Clostridium welchii. J. Hyg, 1953, 51, 102-7.
24PONS JL., PICARD B., NIEL P., LELUAN G., GOULLET P. Esterase electrophoretic polymorphism of human and animal strains of Clostridium perfringens. Appl. Environ. Microbiol., 1993, 59 , 496-501.
25ROOD JI., COLE ST. Molecular genetics and pathogenesis of Clostridium perfringens. Microbiol. Rev., 1991, 55, 621-48.
26ROOD JI., MC CLANE BA., SONGER G., TITBALL RW. The clostridia molecular biology and pathogenesis. London: Academic Press, 1997: 533.
27SALAM ISA., SANOUSI SME. Proposed scheme for isolation and identification of Clostridium perfringens and Clostridium perfringens like organisms. Rev. Élev. Méd. Vét. Pays trop, 1991, 44, 153-8.
28SMITH L. Virulence factors of Clostridium perfringens Rev. Infect. Dis., 1979, 1, 254-60.
29SONGER JG. Clostridial enteric diseases of domestic animals. Clin. Microbiol. Rev., 1995, 9, 1-19.
30SORENSEN K., BRADBECK U. A sensitive protein assay method using micro-titer plates. Experientia, 1986, 32, 161-2.
31STERNE M., BATTY I. Pathogenic Clostridia. London: Butterworths, 1975:.143.
32STRINGER MF., SHAH N., GILBERT RJ. Serological typing of Clostridium perfringens and its epidemiological significance in the investigation of food poisoning outbreaks. In: INTERNATIONAL SYMPOSIUM ON FOOD MICROBIOLOGY AND HYGIENE, 10, Szczecin, 1979. Proceedings... Szczecin: Poland. 1979. 189.
33SUTTON JC., WILLIAMS PH. Comparison of extracellular polysaccharides of Xanthomonas campestris from culture and from infected cabbage leaves. Can. J. Bot., 1970, 48, 645-51.
34TSAI C., TORRES-ANJEL MJ., RIEMANN HP. Improved culture techniques and sporulation medium for enterotoxin production by C. perfringens type A. J. Formos. Med. Assoc., 1974, 73, 404-9.
35TSCHIRDEWAHN B., NOTERMANS S., WERNARS K., UNTERMAN F. The presence of enterotoxigenic C. perfringens strains in faeces of various animals. Int. J. Food Microbiol., 1991, 14, 175-8.
36UEMURA T., NAKAHARA H., HORIGUCHI Y., SAKAGUCHI G. Trypsinization of derivatives of C. perfringens enterotoxin. Lett. Appl. Microbiol., 1986, 3, 31-3.
37VOLLER A., BIDWELL DE., BARTLET A. Enzyme immunoassays in diagnostic medicine. Bull. World Health Organ., 1976, 53, 55-6.
38WATSON GN. The assessment and application of a bacteriocin typing scheme for Clostridium perfringens J. Hyg., 1985, 94, 69-79.
39WILLIS AT. Clostridia of wound infection London: Butterworths, 1969: 41-156.
CORRESPONDENCE TO:L. Baldassi Laboratório de Bacteriologia Geral e Micobacterioses, Instituto Biológico. Avenida Conselheiro Rodrigues Alves, 1252, 04014-002, São Paulo, SP, Brasil.E-mail:
Publication in this collection
13 May 2002
Date of issue
21 June 2001
05 Mar 2001