Antimicrobial activities , toxinogenic potential and sensitivity to antibiotics of Bacillus strains isolated from Mbuja , an Hibiscus sabdariffa fermented seeds from Cameroon

Department of Food Engineering and Quality Control, University Institute of Technology, P. O. Box 454 Ngaoundere, Cameroon. Department of Food Science and Nutrition, National School of Agro-Industrial Sciences, P. O. Box 455 Ngaoundere, Cameroon. Laboratoire Universitaire de Biodiversité et Ecologie Microbienne (EA 3882), Université de Brest, Université Européenne de Bretagne, ESIAB, Technopôle Brest-Iroise, 29280 Plouzané, France. Laboratoire de Microbiologie des Environnements Extrêmes (LMEE), Université de Brest, Université Européenne de Bretagne, IUEM-UMR 6197 CNRS-IFREMER, Place Nicolas Copernic, F-29280 Plouzané, France.

Author(s) agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License Mbuja is commonly used in the diet of people from the Sahelian region, especially for those living in rural areas with low incomes.The condiment is mostly appreciated for its nutritional value and organoleptic properties and it is used to flavour soups and sauces (Mohamadou et al., 2009).Earlier studies revealed that Mbuja could contribute to consumers' health through its anti-oxidants and phenolic compounds (Mohamadou et al., 2007).Bacillus spp.mainly Bacillus subtilis and related genera are responsible for the fermentation (Mohamadou et al., 2013).
Like for many other traditional condiments, the production of Mbuja relies on spontaneous and uncontrolled fermentation that could impact the quality and safety of the products on one hand, and on the consumers' health on the other hand.Recent studies brought evidence that safety of fermented products could be significantly improved by selecting starter cultures with protective effects against most common pathogenic and spoilage bacteria, yeasts and molds (N'dir et al., 1994;Ouoba et al., 2007).Some Bacillus species produce antibiotics and antimicrobial compounds such as bacteriocins including subtilin and subtilosin synthesized by B. subtilis (Klein et al., 1992;Stein et al., 2004;Abriouel et al., 2011).On the other hand, Bacillus cereus and related species (Bacillus thuringiensis and Bacillus anthracis) are well-known food poisoning bacteria which produce either emetic heat-stable toxin or diarrheal enterotoxins (HBL, NHE and BcET) (Matarante et al., 2004).Another trait of Bacillus cereus and related species toxicity is their ability to secrete phospholipases, cell-lysing enzymes, including sphingomyelinase (sph) and phosphatidylinositol-and phosphatidylcholinespecific phospholipase (piplc) (Matarante et al., 2004).Furthermore, some studies reported that virulence factors were also present in other Bacillus species than Bacillus cereus (Kramer and Gilbert, 1989;Phelps and McKillip, 2002).
The aim of this study was to investigate the potential food preservation and contribution to food safety as criteria for selection of starter cultures to be used in controlled fermentation of H. sabdariffa seeds.Hence, the work intended to assess antimicrobial activity, toxinogenic potential and sensitivity to antibiotics of Bacillus strains isolated from Mbuja.

Preparation of Bacillus inocula
Bacillus strains cryopreserved in culture medium containing glycerol at 30% (v/v) were cultured on nutrient agar (NA) (AES Chemunex, Bruz, France) plates and incubated for 24 h at 37°C.The strains were then subcultured for 18 h at 37°C in 10 mL of trypton soy broth (TSB) (AES Chemunex, Bruz, France).These cultures were used for antimicrobial activities screening.

Preparation of fungal inocula
Molds' inocula were prepared by growing the molds on potato dextrose agar (PDA, AES Chemunex) slants at 25°C for 7 to 10 days until sporulation.The spores were then collected by vigorously shaking the slants with sterile peptone water (0.1%, w/v).Yeast cell inocula were prepared from cultures grown in M2Lev agar (20 g/L malt extract, 3 g/L yeast extract, and 15 g/L agar).Spores and yeasts concentrations were determined by microscopic Malassez cell counts and adjusted to 10 6 cells or spores/mL with sterile peptone water (0.1%).

Inhibition of indicators bacteria by cross-streak assay
The screening of inhibitory activity of Bacillus strains was realized according to the cross-streak assay described by Pugsley and Oudega (1987) with some modifications.Each Bacillus strain was picked from TSB and a single streak of this culture was applied both on a NA agar plate and on a glucose agar with bromocresol purple (BCP) plate using a 10 µL sterile loop.Cross-streak assay on BCP was aimed at verifying that inhibitory activity of Bacillus strains was not due to acid production.After incubating the plates for 18 h at 37°C, the densely overgrown streak was inactivated by applying chloroform (anhydrous, ≥ 99%, Fluka, Sigma-Aldrich, Saint-Quentin Fallavier, France) using a Pasteur-pipette to form a thin film covering the whole streak of bacteria.The Petri dish was left closed for 10 min to inactivate all living cells before they were opened for 10 min (under a hood) to evaporate the toxic chloroform.A streak of the tested pathogenic bacteria was then applied perpendicularly to the chloroform-inactivated Bacillus strain (Figure 1) to determine the inhibition activity.Standardized suspensions in 0.85% NaCl (Mc Farland Standard 1) of tested pathogen was applied with a 10 µL sterile loop as indicated by Zihler et al. (2009).

Antifungal activity screening
The antifungal assay was performed by the overlay method described by Magnusson and Schnürer (2001).The method was realized using NA plates on which Bacillus strains were inoculated as a spot and incubated at 30°C.The plates were then overlaid with 10 mL of M2Lev soft agar (0.8% agar) containing 10 6 yeast cells or fungal spores per mL.The plates were examined qualitatively for clear zones of inhibition around the bacterial spots.

Total bacterial DNA extraction
The 26 Bacillus strains were grown for 18 h at 30°C on TSB (AES Chemunex).Tubes were centrifuged 10 min at 6000 g and the pellets were suspended in 0.5 mL of sterile physiological water (0.85% NaCl).DNA was extracted from suspended pellet and purified using the FastDNA ® SPIN Kit (QBIOgene; MP Biomedicals, Solon, Ohio, USA) as recommended by the manufacturer's SPIN TM protocol.Pure DNA samples were frozen at -20°C until use.

PCR detection of genes encoding the bacteriocins subtilin and subtilosin
PCR were performed to screen for the presence of subtilin and subtilosin genes, two bacteriocins frequently encountered in Bacillus species (Abriouel et al., 2011).All the Bacillus strains were tested.Primers designed by Sutyak et al. (2008) were used for targeting the subtilin gene (spaS), whereas for subtilosin A (sboA), primers were designed using Primer 3 software (Rozen and Skaletsky, 2000) on alignment of known sequence encoding the sboA gene from B. subtilis subsp.spizizenii (acc.n°BSUW23_18455) (Table 1).One microliter (95-105 ng) of genomic DNA from each Bacillus strains or B. subtilis LMG 8197 used as positive control were added to 24 µl of a mix PCR consisting of each primer (1 µM), dNTP (0.2 µM), MgCl 2 (1.5 µM), buffer (5 µL of 5× buffer), water (10 µL) and 0.5 µL of GoTaq polymerase (Promega, Charbonnières, France) to make a final volume of 25 µL.PCR was carried out using a PTC-100 programmable thermal controller (MJ Research, Waltham, Massachussetts, USA) according to the following conditions as described by Sutyak et al. (2008): denaturation for 30 s at 94°C, annealing for 30 s at 55°C (spaS) or 50°C (sboA) and elongation for 1 min at 65°C for a total of 30 cycles.PCR products were electrophoresed on agarose gel (1.5% w/v), stained with ethidium bromide and visualized under UV.

Bacillus thuringiensis strains
The presence of Cry1 genes commonly encountered in B. thuringiensis and known to encode for proteins active against many insect orders including Lepidoptera, Diptera, Coleoptera, Hymenoptera, Homoptera, Mallophaga, and Acari (Cinar et al., 2007) was also investigated in the two B. thuringiensis strains (S10 and SAc) with primers (CJI-1 and CJI-2) designed by Céron et al. (1995) (Table 1).PCR mixtures were prepared using 1 µl of each primer (25 µM), 0.5 µl of dNTP (10 µM), 1.5 µl of MgCl 2 (25 µM), 5µl of 5× Green Buffer GoTaq® Flexi Buffer, 0.5 µl of GoTaq® DNA Polymerase (Promega), and 14.5 µl of water to make a final volume of 24 µL.Amplification was performed using 1 µl (about 100 ng) of genomic DNA from Bacillus thuringiensis strains and B. subtilis LMG 8197 (negative control) and 24 µl of PCR mixture in a PTC-100 programmable thermal controller (MJ Research) according to the following conditions defined by Bobrowski et al. (2001): initial denaturation at 95°C for 2 min followed by 30 cycles consisting of a denaturation step of 95°C for 1 min; annealing step of 52°C for 1 min, and an extension step of 72°C for 1 min and a final extension of 72°C for 5 min.PCR products were analyzed by electrophoresis, stained with ethidium bromide and visualized under UV.

PCR screening of genes encoding B. cereus enterotoxins and virulence factors
The genes encoding the major B. cereus enterotoxins and virulence factors were investigated by PCR in the 26 Bacillus strains.Parts of the following genes encoding hemolysin (hbl-D/A, 623 bp), non hemolytic enterotoxin (nheB, 769 bp), B. cereus enterotoxin T (bceT, 428 bp) and enterotoxin FM (entFM, 1269 bp) were screened.Virulence factors were also investigated with the targeting of genes coding for two phospholipases associated with cell lysis, sphingomyelinase (sph gene, 558 bp) and phosphatidylinositol-specific phospholipase C (piplc gene, 569 bp).Reference studies and used primers (Sigma-Aldrich) are listed in Table 1.PCR mixture of 25 µl contained 20 to 35 ng of genomic DNA (Bacillus strains and Bacillus cereus LMG 6923 used as positive control), 1 µM of each primer, 0.5 µM dNTP, 1.5 µM of MgCl 2 , GoTaq® DNA Polymerase, 5× Green Buffer GoTaq® Flexi Buffer and distilled water.Amplification was realized according to Matarante et al. (2004) and consisted of an initial denaturation at 94°C for 3 min, followed by 40 cycles of 94°C for 25 s, 55°C for 45 s, and 72°C for 2 min and a final extension at 72°C for 5min.PCR was performed with a PTC-100 programmable thermal controller (MJ Research).PCR products were analyzed on 1.5% (w/v) agarose gel stained with ethidium bromide and visualized by UV.

Hemolytic activity of Bacillus strains
Hemolytic activity was determined for all strains on blood agar plates (AES Chemunex) containing 5% of sheep blood.B. subtilis LMG 8197 and Bacillus cereus LMG 6923 were used respectively as negative and positive controls.Five microliters of cells grown overnight in TSB were inoculated as spots on blood agar plates which were incubated at 30°C for 24 h.Positive strains produced clear zone of hemolysis around the colonies.

Antibiotic susceptibility of Bacillus strains
The antibiotic susceptibility screening of the Bacillus strains was carried out by the disc diffusion method (Biomérieux, Marcy l'Etoile, France) as indicated by the producer.A single colony of Bacillus strains grown on NA plate was diluted in sterile 0.85% NaCl

Antimicrobial activities of the Bacillus strains
The ability of the 26 Bacillus strains to inhibit indicators and pathogenic bacteria, yeasts and molds was investigated and reported in Tables 3 and 4. The antibacterial behaviour of Bacillus strains strongly varied between Bacillus strains and targets.Bacillus strains were both effective against Gram-positive and Gramnegative bacteria.Bacillus cereus ATCC6464 was inhibited by 14 strains, Listeria innocua HPB13 by 10 and S. aureus by 2 strains only.E. coli was the most sensitive -, no inhibition; +, zone of inhibition between 6 and 8 mm; ++, zone of inhibition higher than 8 mm.
Gram-negative bacteria to Bacillus strains with nine inhibitions whereas P. aeruginosa PAO1 was inhibited with one strain only and S. typhimurium was not inhibited.

PCR screening for bacteriocins and larvicidal proteins genes
The 26 Bacillus strains were assayed for the presence of

PCR screening for toxins encoding genes
The presence of virulence genes was assessed for the 26 Bacillus strains by PCR screening (Figures 2 and 3).
The presence of some genes of the enterotoxins responsible for diarrheal poisoning was confirmed for two strains identified as B. thuringiensis.Indeed, non hemolytic enterotoxin (nhE) genes were amplified by the specific primers in B. thuringiensis S10 and SAc while hemolysin gene (hbl-D/A) was detected only in SAc.However, Bacillus cereus enterotoxin (BcET) and enterotoxin FM genes were not amplified in all strains including the two B. thuringiensis.Two Bacillus cereus enzymes encoding genes were also screened: sphingomyelinase (Sph) and Phosphatidylinositol-specific phospholipase C (Piplc).These haemolytic and cell membrane hydrolytic enzymes genes were amplified in the 2 B. thuringiensis S10 and SAc.No other Bacillus species contained Sph and Piplc genes.

Hemolytic activity
Hemolytic activity was tested for the 26 Bacillus strains.
Only the two B. thuringiensis strains (S10 and SAc) exhibited haemolytic activity on 5% sheep blood agar plates.

Antibiotics susceptibility
Relative susceptibilities to eight antibiotics were determined for 26 Bacillus strains (Table 5).The most important antibiotic resistances for all species were observed in chloramphenicol.The most susceptible species were B. megaterium with resistance (40%) to only one antibiotic (chloramphenicol).This study was carried out to investigate the potential food preservation and contribution to food safety as criteria for selection of starter cultures to be used in controlled fermentation of Hibiscus sabdariffa seeds.Both bacteria and fungi (yeasts and molds) are important food spoilage or pathogens in different food systems in developing countries.Some microorganisms evaluated in this study are regularly cited in food toxiinfections cases in Cameroon.Therefore, there is a need for affordable and safe methods to inhibit bacterial and fungal growth in fermented foods.This work documents the control of food spoilage and pathogenic bacteria and fungi by 26 Bacillus strains belonging to 5 species and their potential use as starters with a protective potential for a controlled and safe fermentation process.The strains showed different antibacterial and antifungal profiles between and within identified species.Bacillus strains active against bacteria were not effective against fungi.
With regard to antibacterial activities, B. subtilis and related species (B.amyloliquefaciens) induced different degree of inhibition depending on the tested species, independently of their Gram staining.Differences in antibacterial properties observed between closely related species were also observed in their phenotypic characteristics and could be explained by their genetic diversity (Mohamadou et al., 2013).The most antibacterial species were B. amyloliquefaciens followed by B. subtilis whereas B. megaterium induced no inhibition of bacteria and B. pumilus only two.Bacillus strains mainly inhibited, in a decreasing order, B. cereus, L. innocua (used as a model for L. monocytogenes), S. aureus and E. coli, which are among the most common foodborne pathogenic bacteria in Cameroon (Njongmeta et al., 2004;Djouldé et al., 2007).The inactivation tests showed that B. amyloliquefaciens S1 and S5 and B. subtilis S12 exhibited the broadest activity against the most common pathogens studied.B. amyloliquefaciens S1 and S5 inhibited Gram positive (B.cereus ATCC6464, L. innocua HPB13 and S. aureus) and Gram negative (E. coli and P. aeruginosa PAO1) bacteria.However, that none of the 26 strains was able to inhibit Salmonella Typhimurium, a leading and endemic cause of bloodstream infection in sub-Saharan Africa (Morpeth et al., 2009) is disappointing.
The pathways of antibacterial activities were partially screened.The cross-streak test on BCP indicated that the Bacillus species did not inhibit bacteria indicators through acid production, because the medium did not turn yellow as a sign of acidification.PCR screening of spaS and sboA genes showed that seven B. subtilis strains were susceptible to produce known bacteriocins: subtilin and/or subtilosin A. The production of these 2 antibacterial peptides is well documented and has been reported for B. subtilis and other related species like B. amyloliquefaciens (Klein et al., 1992;Stein et al., 2004;Sutyak et al., 2008).Subtilin, a cationic pentacyclic antimicrobial peptide is a lantibiotic that shows antimicrobial activity against a broad spectrum of Grampositive bacteria (Abriouel et al., 2011).Subtilosin A shows a bactericidal activity against Gram-positive and Gram-negative bacteria (Shelburne et al., 2007;Sutyak et al., 2008).However, none of the most antibacterial strains (S1, S5 and S12) appears equipped to produce neither subtilin nor subtilosin A. These results suggest that B. subtilis and B. amyloliquefaciens strains isolated from Mbuja probably produced other antibacterial molecules active against Gram-positive and Gramnegative bacteria or they possess divergent sequences.
Molds and yeasts may also be pathogenic (through their toxins, like aflatoxins) and important food spoilage organisms in Cameroon (Mbiapo et al., 1989;Djouldé et al., 2007).Bacillus isolates were able to inhibit in vitro fungal growth.Both molds and yeasts were sensitive to representatives of the studied species.The most important antifungal activities were reported for B. megaterium strains while B. subtilis proved to be very weak fungi inhibitors.The strain B. megaterium S8 showed the broadest antifungal spectrum but inhibited mostly yeasts.In contrast, B. amyloliquefaciens S5 inactivated only two fungi, A. flavus and F. oxysporum, but they are among the most common food spoilage and poisoning fungi encountered in developing countries (Yaouba et al., 2010).
As said earlier, none of the Bacillus strains produced sufficient organic acid to inhibit studied indicators.In addition, most fungi are weakly sensitive to organic acid.That B. amyloliquefaciens inhibits fungi is not surprising since Yoshida et al. (2001) earlier reported antibacterial and antifungal activity of this species through bacteriocinlike inhibitory substances (BLIS).But to our knowledge, the available literature has not yet reported the antifungal activity of B. megaterium, mostly known as a producer of broad spectrum bacteriocin active against food spoilage bacteria (Khalil et al., 2009).This study describes for the first time strains of B. megaterium active against fungi but not against bacteria.
Although no history of food intoxication by Mbuja was recorded (Mohamadou et al., 2009), the presence of at least one member of the B. cereus group (a well known poisoning organism) and the rare but possible production of toxin by non-cereus strains (From et al., 2005)  This study suggests that B. thuringiensis strains S10 and SAc should be excluded in starter formulation to produce a controlled Mbuja.Nevertheless, another interesting use of B. thuringiensis is their potential to protect food crops from insects, pending on their ability to produce Cry1 proteins.These proteins are insecticidal on certain insects but not toxic to other insect, plants and animal.Due to its selective and specific action, the Cry1 protein has been use as efficient biological insecticide and an alternative to chemical insecticide (Bobrowski et al., 2001).B. thuringiensis tested in the present work were analysed for the presence of Cry1 gene.No amplification product could be detected by PCR for these 2 strains.Recently, Bozlagan et al. (2010) investigated the presence of Cry1 gene in 60 B. thuringiensis strains from agricultural fields and their bioactivity against larvae.These authors showed that only 17 isolates carried the Cry1 gene, indicating that some B. thuringiensis may not have this gene.However, the absence of Cry1 gene is not indicative of the total absence of insecticidal activity (Bobrowski et al., 2001).
Despite the absence of toxin genes in most of the species (except for B. thuringiensis), the interest in antibiotic resistance of the Bacillus strains could be justified by their possible side effects on certain Mbuja consumers.Indeed the immune-compromised consumers, in the context of general outbreak of HIV/AIDS in developing countries, may face increased risks of opportunistic infections.In addition possible antibiotic resistance gene transfer between Bacillus spp.and the intestinal microbiota on one hand and between Bacillus spp.and pathogenic bacteria on the other hand must be considered.All the strains displayed diversity in their susceptibility and resistance to the 8 antibiotics tested.The most effective antibiotics were erythromycin, vancomycin and streptomycin.Most species were resistant to chloramphenicol.Resistance to antibiotics was strain-dependent.Hence, B. subtilis, the main fermenting species isolated in Mbuja, exhibited 18% to 64% resistance to tetracyclin and chloramphenicol while all B. thuringiensis were resistant to ampicillin and chloramphenicol.Similar diverse susceptibilities were reported for these species and their close relatives by other studies in foods, environmental and clinical samples (Aslim et al., 2002;Schlegelova et al., 2003;Dautle et al., 2004;Luna et al., 2007;Adewumi et al., 2009;Chaves et al., 2011).At the whole, most of the strains were sensitive to different classes of antibiotics.
The most active Bacillus spp.against pathogenic and food spoilage micro-organisms were either antibacterial or antifungal.They were totally exempted with toxins genes and were sensitive to different antibiotics.A previous paper (Mohamadou et al., 2007) stated that B. amyloliquefaciens (S1 and S5) were highly amylolytic and proteolytic while B. subtilis S12 and B. megaterium S8 and S9 were highly to moderately proteolytic.These traits are important technological properties in fermenting proteinaceous Hibiscus sabdariffa seeds to produce Mbuja.These strains could therefore present a double significant advantage on the nutritional value and safety of Mbuja.
This work was carried out to contribute in selecting starter cultures for controlled production of Mbuja by assessing safety and potential risk of Bacillus strains.In the development of the starter culture, strains with important antimicrobial activities and sensitive to a maximum of antibiotics should be encouraged while toxinogenic and antibiotic-resistant strains should be discouraged.In this respect, this study raised the interest of five strains.B. subtilis S12 and B. amyloliquefaciens S1 were not toxinogenic, active against most pathogenic bacteria tested and were sensitive to all antibiotics.B. megaterium S9 was also not toxinogenic, sensitive to all antibiotics and active against 5 fungi.B. amyloliquefaciens S5 with strong activity against bacteria and two important pathogenic molds should also be considered.Fortunately, these strains showed good technological properties.It will be suitable to undertake Hibiscus sabdariffa seeds fermentations based on the combination of these species to test their technological as well as their protective properties against a wider panel of pathogenic strains within Mbuja.

Figure 1 .
Figure 1.In vitro activity of Bacillus strains by Cross-streak with strong inhibition (A) and no inhibition on BCP agar (B).

Figure 2 .Figure 3 .
Figure 2. PCR gel of hemolysin hbl D/A (623 pb) and Enterotoxin FM entFM (1269 pb) genes.R is the reference strain of Bacillus cereus LMG 6923 used as a positive control, and B1 is the reference strain of Bacillus subtilis LMG 8197 used as a negative control.
two bacteriocins (subtilin and subtilosin) genes generally present in B. subtilis and related species.The PCR investigation revealed the presence of spaS (subtilin gene) in B. subtilis S2 only.Subtilosin A gene (sboA) was detected in B. subtilis S2, S7, S15, S17, S18, S19 and S20.Seven strains out of 11 B. subtilis contained either spaS or sboA.The screening of these bacteriocins encoding genes was negative for all other Bacillus species.The Cry1 gene, encoding for a protein (Cry1) active against many insects was not detected in any of the 26 tested Bacillus strains including B. thuringiensis S10 and SAc.

Table 1 .
Primers used for PCR detection of genes encoding toxins, virulence factors, antimicrobial peptides and insecticidal proteins.

Table 2 .
Interpretive criteria for MIC (mg/L) and MID (mm) by disc diffusion method.
bacterial film.The plates were then left to dry for 10 min before the disc were applied in the center of the plate.The plates were incubated at 37°C for 48 h.For tests strains susceptible to antibiotics, a clear area was observed around the disc.The diameters of inhibition areas were measured and interpreted according to the recommendations of the Antibiogram Committee of SFM, the French Society for Microbiology (2009) (Table2).

Table 3 .
Antimicrobial activity of the 26 Bacillus strains obtained with the cross-streak assay against 6 indicator bacteria.

Table 4 .
The broadest spectrum of inhibition was shown by B. Antimicrobial activity of the 26 Bacillus strains obtained with the spot test assay against 9 indicator fungi.
amyloliquefaciens S1 and B. subtilis S12 (with 4 sensitive targets) while the strongest effects were observed with B. amyloliquefaciens S5.Bacillus cereus showed weak inhibition ability and no B. megaterium was able to inhibit the tested bacteria.As for bacteria, the ability of tested Bacillus spp. to inhibit fungal targets varied with Bacillus strains and tested fungi.However, the antifungal Bacillus species (B.megaterium, B. thuringiensis and B. pumilus) differed from the antibacterial ones (B.amyloliquefaciens and B. subtilis) (Table 4).Both molds and yeasts were inhibited.The most sensitive molds were Fusarium oxysporum UBOCC-A-108079 (sensitive to 9 Bacillus isolates), followed by Aspergillus flavus UBOCC-A-10826 and Mucor plumbeus CBS129.41(3 Bacillus) while Debaromyces hansenii CLIB197 and Kluyveromyces

Table 5 .
Antibiotic susceptibility (percentage) of the 26 Bacillus strains against nine tested antibiotics.
led us to investigate the toxinogenic potential of the 26 isolates.Virulence factors and toxin encoding genes usually present in B. cereus were not detected in 24 strains of B. subtilis, B. amyloliquefaciens, B. megaterium and B. pumilus.However, positive PCR results were obtained for hbL, nhE, spH and Piplc genes but not for bceT and entFM genes on the two B. thuringiensis strains tested.These interesting results confirm the very low risk of food-borne disease due to other Bacillus species than B. cereus closest relatives and support their use as starter in a controlled fermentation process to produce a safe Mbuja.The present results go in the same line as those obtained by Matarante et al. (2004) who reported the absence of B. cereus toxin encoding genes in B. subtilis and B. pumilus isolated in industrial and artisanal cured sausages in Italy.When haemolytic activity was tested, only the two B. thuringiensis produced halos indicating a strong haemolytic power common in B. cereus and close related species.