Common occurrence of enterotoxin genes and enterotoxicity in Bacillus thuringiensis


  • Adelaida M. Gaviria Rivera,

    1. Department of Biological Sciences, Heriot-Watt University, Riccarton, Edinburgh EH14 4AS, UK
    Search for more papers by this author
  • Per Einar Granum,

    1. Department of Pharmacology, Microbiology and Food Hygiene, Norwegian School of Veterinary Science, N-0033 Oslo, Norway
    Search for more papers by this author
  • Fergus G. Priest

    Corresponding author
    1. Department of Biological Sciences, Heriot-Watt University, Riccarton, Edinburgh EH14 4AS, UK
      *Corresponding author. Tel.: +44 (131) 451 3464; Fax: +44 (131) 451 3009, E-mail address:
    Search for more papers by this author

*Corresponding author. Tel.: +44 (131) 451 3464; Fax: +44 (131) 451 3009, E-mail address:


Seventy-four strains of Bacillus thuringiensis representing 24 serovars were examined for the presence of three enterotoxin genes/operons; the non-haemolytic enterotoxin Nhe, the haemolytic enterotoxin hbl and the Bacillus cereus toxin bceT using polymerase chain reaction. The nheBC genes were found in all strains examined, the hblCD genes in 65 of the 74 strains and bceT in 63 strains. There was little consistency of the distribution of enterotoxin loci among strains of the same serovar in serovars that were well represented in our collection. Culture supernatants from all but one strain inhibited protein synthesis in Vero cells, generally with a toxicity equivalent to that seen in strains of B. cereus isolated from incidents of food poisoning.


Bacillus thuringiensis is a typical endospore-forming bacterium distinguished by the accumulation of polypeptides forming a crystalline parasporal inclusion during sporulation [1]. These crystal proteins may be toxic to insects, leading to the extensive use of B. thuringiensis as biological insecticides for crop protection [1,2]. Bacillus cereus is a common soil inhabitant that is often implicated in food poisoning, particularly in Scandinavia and Canada [3,4]. It is now apparent that, in addition to phenotypic similarity, B. cereus and B. thuringiensis strains have virtually identical 16S and 23S rRNA genes [5,6], and such similar genomic composition that they are often considered to be members of the same species [7,8,9].

B. cereus is responsible for two forms of food poisoning. The emetic syndrome is induced by a cyclic peptide that is released into foods, especially rice, during growth of appropriate strains of the bacterium and causes vomiting some 2 to 6 h after ingestion [10,11,12]. The diarrhoeal form of the disease is caused by the secretion of enterotoxins by B. cereus cells in the intestine following consumption of contaminated foods [13]. Several enterotoxins have been implicated in this process of which two tripartite protein complexes, the haemolytic (Hbl) toxin [14,15] and the non-haemolytic (Nhe) toxin [16,17], are the most important. Additionally, a single component toxin (BceT) exhibits Vero cell cytotoxicity and induces fluid accumulation in ligated ileal loops [18], but is of unknown contribution to the human disease. Given the taxonomic similarity of B. cereus and B. thuringiensis strains and the introduction of high numbers of B. thuringiensis spores into the human food chain through vegetables treated with B. thuringiensis-based insecticides, we examined numerous strains of B. thuringiensis representing 24 serovars for the presence of hbl, nhe and bceT genes and cytotoxicity. Here we show that enterotoxin genes are common in B. thuringiensis strains and that the toxicity of culture supernatants is generally at the same levels as those associated with food poisoning strains of B. cereus.

2Materials and methods

2.1Strains and growth conditions

The strains of B. thuringiensis were obtained from the Collection of Bacillus thuringiensis and Bacillus sphaericus, Institut Pasteur, Paris (strains labelled with a ‘T’ in Table 1) or were from our own collection. Strains were grown routinely on NSM agar (nutrient agar supplemented with 5×10−5 M MnCl2, 7×10−4 M CaCl2 and 1×10−3 M MgCl2) and in nutrient broth containing 0.1% glucose for preparation of cells for DNA extraction. Strains were maintained as suspensions of spores and cells at −20°C in 20% (v/v) glycerol.

Table 1.  Occurrence of enterotoxin genes and cytotoxicity in some strains of B. thuringiensis
  1. a+, amplification of PCR product of correct size; N, no amplification product.

  2. b+++, between 90 and 100% inhibition (100 μl cell extract); ++, between 50 and 90% inhibition; +, less than 50% inhibition; ND, not done.

T01001, T01022, T013261thuringiensisN+++++
T03A001, T03A075, T03A075, T03A172, T03A2873a, 3b, 3ckurstaki++++++
T03A3613a. 3b, 3ckurstaki+N++++
T03C0013a, 3d, 3efukuokaensis+N++++
T04002, T04016, T040244a, 4bsotto++++++
T042364a, 4bsotto+++ND
T04A0174a, 4bsotto biotype dendrolimusNN++++
T04A023, T04A0344a, 4bsotto biotype dendrolimus++++++
T04A0614a, 4bsotto biotype dendrolimus+N++++
T04B001, T04B012, T04B054, T04B060, T04B0734a, 4ckenyae++++++
T05A001, T05A015, T05A0195a, 5ccanadensis++++++
T05A0305a, 5ccanadensisN+++++
T06002, T06007, T06010, T060246entomocidus++++++
T06A0066entomocidus biotype subtoxicus+++ND
T07058, T07146, T07180, T071967aizawai++++++
T08001, T08009, T08012, T08023, T080318a, 8bmorrisoni++++++
T080258a, 8bmorrisoni+++ND
T09010, T09011, T090249tolworthi++++++
T10001, T10003, T10016, T1001810a, 10bdarmastadiensis++++++
T1001710a, 10bdarmstadiensis+++ND
T1002410a, 10bdarmstadiensis+N++++
T11A001, T11A00211a, 11ckyushuensis++++++
T13001, T13004, T1302813pakistani++++++
T20001, T2000220a, 20byunnanensis++++++
T28A00128a, 28cjeganthesan++++++

2.2Screening strains for the presence of enterotoxin genes

Chromosomal DNA was prepared using the Puregene DNA isolation kit (Philip Harris/Flowgen, Shenstone, UK) from 5 ml cultures of B. thuringiensis grown for 18 h at 37°C. Primers for polymerase chain reaction (PCR) amplification of the enterotoxin genes were: bceT, BceTF (position 1173–1192), 5′-GAG TTA GTT TCA ACA GCG CG and BceTR (position 1746–1765), 5′-TCA GGC TGA TCT AGT AGA CC; hbl, HblCF1 (position 1446–1465), 5′-GGA ATG GTC ATC GGA ACT CT and HblDR1 (position 3501–3520) and nhe, NheBF1 (position 1741–1760), 5′-ATG TAT CGT CTG TTG ATG CG and NheCR1 (position 2953–2972), 5′-TAC TTG ATG TCG TTT GTG CC. PCR reactions were conducted in 50-μl volumes using an initial denaturation at 95°C for 10 min followed by 30 cycles of denaturation at 95°C for 30 s, annealing for 30 s at 56°C for bceT, 53°C for hbl and 51°C for nhe and extension at 72°C for 2 min. The final extension time was increased to 10 min. PCR products were electrophoresed in 1% agarose gels in Tris acetate buffer and visualised by staining with ethidium bromide. Where faint or indefinite bands were recovered, PCR amplifications were repeated with new DNA preparations.

2.3Southern hybridisation

Chromosomal DNA was digested with HindIII, electrophoresed and transferred to nylon membranes as described previously [19]. Digoxigenin-labelled probes were prepared by PCR using the nhe and hbl primers with the reaction supplemented with DIG-11-dUTP according to the manufacturer's instructions (Boehringer). Hybrids were visualised with alkaline phosphatase and X-phosphate.

2.4Cytotoxicity of culture fluid

Strains were grown in brain heart infusion supplemented with 1% glucose at 32°C to late exponential phase (6 h) and harvested by centrifugation. Culture supernatant was examined for the inhibition of 14C-leucine uptake in Vero cells [20] as described previously [17]. For each strain, undiluted culture supernatant (100 μl) was incubated with Vero cells and counts incorporated measured after 120 min. For strain T26001, it was necessary to concentrate the supernatant fluid 10-fold by protein precipitation with ammonium sulfate to 80% saturation (56 g l−1). After recovery of the protein by centrifugation (10 000×g for 20 min) the pellet was resuspended in 20 mM phosphate buffer at pH 6.8 in about 1/20 of the original volume. The remaining ammonium sulfate was removed by dialysis against the same buffer at 4°C for at least 6 h and the volume then adjusted to one tenth of the original volume. Results are given semi-quantitatively (see Table 1 and Fig. 3).

Figure 3.

Inhibition of protein synthesis (as reduction in 14C-leucine uptake) in Vero cells by extracellular protein extracts form three strains of B. thuringiensis; (◯) T01246, nhe+, hbl+, bceT+; (□) T01015, nhe+, hbl, bceT; (▵) T26001, nhe+, hbl, bceT. Food-poisoning strains of B. cereus generally show >80% inhibition of protein synthesis in Vero cells using 50 μl cell extract (unpublished data from, Norwegian reference laboratory for B. cereus).


3.1Occurrence of enterotoxin genes in B. thuringiensis

PCR amplifications from chromosomal DNA produced the anticipated products of about 1.2 kb for nheBC, 1.9 kb for hblCD and 600 bp for bceT (Fig. 1A,B,C). Examination of the strain collection revealed that nhe was the most common enterotoxin with the nheB and C genes detected in every strain examined (Table 1). hblC and D genes were present in 65 of the 74 strains examined and the bceT gene was present in 63 strains.

Figure 1.

PCR amplification of enterotoxin genes from some strains of B. thuringiensis. (A) nheBC: lane 1, T04002; lane 2, TO4006; lane 3, T04024; lane 4, T04236; lane 5, T04A017; lane 6, T04A023; lane 7, T04A034; lane 8, T04A061; lane 9, T04B001; lane 10, T04B002; lane T04B054. (B) hblCD: lane 1, T04B060; lane 2, T04B073; lane 3, T05A001; lane 4, T05A015; lane 5, T05A019; lane 6, T05A030; lane 7, T06002; lane 8, T06007; lane 9, T06010; lane 10, T06024; lane 11, T06A006. (C) bceT: as for panel (A).

Strains within some serovars, such as aizawai, entomocidus, kurstaki and thuringiensis revealed an inconsistent pattern of toxin genes but in other instances, such as serovars kenyae and morrisoni all strains had the same full complement of toxin genes. We confirmed the presence/absence of toxin genes in selected strains by Southern hybridisation. HindIII digestions of chromosomal DNA showed the presence of nhhB and C on fragments of DNA ranging from about 1.6 kb (strains T03A001, T03A006 and T06A006) to about 14 kb (strain T11A002). There was conservation within strains from the same serovar and between the serotypes 3 and 6 strains (Fig. 2A). Similar blots with the hblCD probe gave complex patterns (Fig. 2B) with the expected fragment of around 5 kb clearly evident in serotypes 3 and 6 strains but accompanied by extra bands of about 3.5 kb in the former. The lack of hybridising material in strains T05A030 and T10024 confirmed the negative PCR results for these strains.

Figure 2.

Southern blots of chromosomal DNA from various strains of B. thuringiensis hybridised to (A) a probe for nheBC and (B) a probe for hblCD. Lane 1, T03A001; lane 2, T03A075; lane 3, T05A030; lane 4, T06A006; lane 5, T10024; lane 6, T11A002 and lane 7, T39001.

3.2Cytotoxicity of culture fluid

All but one of the strains examined inhibited protein synthesis in Vero cells. Typical high toxicity (shown by +++ in Table 1 and as a dilution experiment in Fig. 3) results in >80% inhibition with 50 μl supernatant and is characteristic of B. cereus strains recovered from cases of diarrhoeal food poisoning [16,17]. Most strains showed this level of toxicity, exceptions being one strain of serovar tolworthi (T09034), one strain of serovar tohokuensis (T17001) and the strain of serovar andalouensis (T37001) that had slightly lower toxicity. Only one strain (T26001, serovar silo) was negative for Vero cell inhibition (Table 1 and Fig. 3) and, although it contained nheBC genes, toxin proteins were not evident on Western blots. However, faint protein bands were apparent when culture fluid was concentrated 10-fold (data not shown) suggesting that very little toxin protein was produced under the growth condition used.


The distribution of enterotoxin genes has been reasonably well established in strains of B. cereus but less so in B. thuringiensis. hblA was found in 10 of 23 (43%) B. cereus strains by Prüb et al. [24] and 26 of 50 (52%) strains by Mäntynen and Lindström [21]. These figures are lower than the occurrence of the product of the hblD gene (L2 protein) from B. cereus strains as detected by the Oxoid kit, which was found in 164 of 194 (84%) B. cereus isolates from foods [22] although using monoclonal antibodies to the Hbl complex indicated that only 50% of the B. cereus strains produced the proteins [23]. The common occurrence of hblCD genes in B. thuringiensis reported here (88%) is in keeping with its detection in all eight strains of B. thuringiensis examined in a previous study [24], and indicates that this operon is as common, if not more so, in B. thuringiensis than in B. cereus. There is less published information for nhe[25], but our unpublished studies show it to be present in most B. cereus strains, supported by detection of NheA protein (equivalent of the hblC gene product, protein L1 of Hbl) in culture fluids from 172 of 194 (92%) B. cereus strains examined [22]. It was present in all the B. thuringiensis strains studied here, again indicating that it is more common in B. thuringiensis than in B. cereus. Finally, the distribution of bceT is controversial having been found in all 10 strains of B. cereus included in the original study [18], but apparently absent from 50 strains of B. cereus of various origins [21], or present in about 40% of strains isolated largely from milk [26]. Our studies with B. thuringiensis are consistent with the last mentioned study since we amplified DNA using primers targeted to bceT from 63 of our 74 strains (85%). In summary, it seems that there is little difference between the distribution of enterotoxin genes among strains of B. thuringiensis and B. cereus and, if anything they are more common in the former.

Given the widespread occurrence of hbl, nhe and bceT, it was not surprising that culture fluids from B. thuringiensis strains were almost invariably highly toxic to Vero cells. The only exception was strain T26001 (serovar silo) which, despite the presence of nheAB, showed no inhibition of protein synthesis in Vero cells. The reason for this may be lack of gene expression since we were able to demonstrate only very low levels of NheABC in Western blots (data not shown) and we are investigating this strain further. Most strains showed high toxicity (+++ in Table 1) which is equivalent to the toxicity associated with B. cereus strains isolated from cases of diarrhoeal food poisoning (Fig. 3). Since B. thuringiensis may occur in foods [22,27], it is important that cooked foods for use in salads or other dishes are stored in such a way to prevent spore germination and bacterial growth. Moreover, it is important for the biocontrol industry to appreciate that enterotoxins are being introduced into the human food chain with the application of B. thuringiensis strains to crops, and to consider undertaking the simple expedient of deleting these genes from their commercial strains.


A.M.G.R. thanks the Columbian Institute COLCIENCIAS for a scholarship.