TEXT

Top Abstract Text References

Psychrotrophic (psychrotolerant) bacteria have been recognized as a recurring problem in the refrigerated storage and distribution of perishable food products, which is particularly pertinent to the dairy industry (13, 15). Research into the bacterial spoilage of dairy products has focused predominantly on postpasteurization contaminants, such as Pseudomonas. However, improved processing conditions for milk and other dairy products have decreased interest in these non-heat-resistant contaminants, with more attention being directed towards psychrotrophic sporeformers such as Bacillus (2, 8, 13, 17). It has been estimated that 25% of all shelf life problems associated with conventionally pasteurized milk and cream products in the United States may be linked to this class of thermoduric bacteria, with a large number of the contaminants being psychrotrophic Bacillus cereus and Bacillus mycoides (13). Moreover, in many European countries, where the average storage temperature of milk products is several degrees higher than that in the United States, psychrotrophic Bacillus species, especially B. cereus, have proven to be an even greater problem to the dairy industry and consumers (1, 8, 16).

B. cereus, Bacillus thuringiensis, and B. mycoides are very closely related bacteria that, in the case of the former two species at least, are known to cause food poisoning through the production of enterotoxins (3, 7, 14). Differentiation of this bacterial group, which also includes Bacillus anthracis, the causative agent of anthrax in humans and animals, has been achieved by using a number of biochemical and molecular techniques (9, 18). However, since all psychrotrophic strains of the B. cereus group can proliferate at refrigeration temperatures (i.e., at or below 7°C), there is also a need to distinguish these cold-adapted strains from their mesophilic counterparts. Early detection and quantification of such psychrotrophic bacteria might allow food manufacturers to more accurately predict the shelf life of perishable products and thus reduce the incidence of food poisoning by strains of the B. cereus group.

The detection of psychrotrophic strains of the B. cereus group in dairy products is at present extremely slow (5 to 10 days), due to the inability to discriminate psychrotrophic and mesophilic strains by means other than growth at low temperature. In order to develop a rapid detection method for such bacteria, at least some molecular aspects of psychrotrophic adaptation need to be known. Studies directed at understanding the molecular mechanisms adopted by Bacillus for dealing with growth at low temperature (i.e., cold adaptation) have recently focused on the role played by a family of small polypeptides termed the major cold shock proteins (5, 11). Certain of these proteins, which are widespread in bacteria (4), are significantly induced by a temperature downshift and appear to play a fundamental role in the bacterium's survival at low temperature (6). These proteins were therefore made the focus of our attention, since it was suspected that subtle differences in their regulation and/or structure might contribute to the psychrotrophic growth of strains of the B. cereus group. Prior studies of the family of major cold shock proteins found in B. cereus have shown CspA to be the homologue most significantly induced by cold shock (11); therefore, initial studies were directed at this particular major cold shock protein.

Psychrotrophic and mesophilic B. cereus and B. mycoides organisms were isolated from milk samples obtained from three dairies in southern Germany (G1, G2, and G3 [Table 1]). Strains were initially isolated on B. cereus selective agar (PEMBA; Oxoid) and tested for the ability to grow at or below 7°C in liquid culture (plate count [PC] medium; Merck). Isolates were then individually characterized according to a range of criteria, including colony morphology, biochemical profiling, and phage typing (12). From these data 28 isolates (9 psychrotrophic B. cereus isolates, 16 mesophilic B. cereus isolates, and 3 psychrotrophic B. mycoides isolates) were characterized as different strains. cspA was PCR amplified from each of these 28 strains and additionally from 4 mesophilic B. cereus strains and 4 mesophilic B. thuringiensis strains gained from other sources (i.e., those strains with cspA signatures shown in Table 1). PCR was performed with a Techne Progene automated thermocycler with 0.2-ml thin-walled PCR tubes (Advanced Biotechnologies). Reactions were carried out in 50-µl volumes containing 5 µl of 10× PCR buffer (supplied with Taq DNA polymerase; Eurogentec), 2.0 mM MgCl₂, 50 pmol of each oligonucleotide primer (see below), 0.2 mM each deoxynucleotide triphosphate (dATP, dCTP, dGTP, and dTTP; Eurogentec), 1 U of Taq DNA polymerase (Eurogentec), and a pinhead-sized aliquot of bacteria picked from a PEMBA plate. Bacterial cells were lysed by heating the mixture at 95°C for 5 min. Amplification of cspA was then attempted, with 30 cycles at 95°C for 15 s, 50°C for 30 s, and 72°C for 30 s, followed by a final extension step at 72°C for 2 min. The oligonucleotide primers used to amplify cspA from the above strains were BcAF2, CGA ATT TGA TAA TGT GTG GAT TC, an oligonucleotide that is specific to a 5' noncoding region of cspA from the psychrotrophic B. cereus strain WSBC10201 (11), and CSPU3, CCC GGA TCC GGT TAC GTT A(G/C)C (A/T)GC T(G/T)(G/C) (A/T/C)GG (G/A/T)CC (degeneracies shown in parentheses), a universal major cold shock protein oligonucleotide that hybridizes at the 3' end of many bacterial cspA homologues, including B. cereus (4).

Alignment of cspA gene sequences obtained from each of the above-mentioned psychrotrophic and mesophilic strains (28 sequences in all) showed that, although the sequences are extremely homologous (alignment of complete sequences not shown), there are three distinct nucleotide positions within the coding region of this gene at which the two thermal groups of bacteria vary consistently (positions 4, 9, and 124 of the cspA coding sequence [11]). These nucleotide differences allowed us to design the discriminatory primers BcAPF1 (GAG GAA ATA ATT ATG ACA GTT) and BcAPR1 [CTT (C/T)TT GGC CTT CTT CTA A] (discriminatory signature bases are shown in boldface, with a single degeneracy shown in parentheses), which favor the amplification of cspA sequences from psychrotrophic strains as opposed to those from mesophilic strains (Table 1). PCR performed with the primers at an annealing temperature of 55°C (with the remainder of the PCR conditions identical to those described previously) allowed a 160-bp DNA fragment of cspA to be amplified only from psychrotrophic strains of the B. cereus group. Identical results were achieved with an annealing temperature of 52°C, providing a 3°C margin of error (e.g., allowing for temperature variations between PCR blocks). In addition to these psychrotrophic cspA primers, a third primer was designed to allow the detection of all strains of the B. cereus group (both psychrotrophic and mesophilic strains), providing a positive control to confirm the presence of mesophilic strains that were not amplified by BcAPF1 and BcAPR1. This third primer, designated BcFF2 (GAG ATT TAA ATG AGC TGT AA), was designed from a 5' noncoding region found in both psychrotrophic and mesophilic B. cereus cspF genes (our unpublished data). Due to the high degree of similarity found between major cold shock protein homologues (11), BcAPR1 is able to hybridize not only to psychrotrophic cspA sequences but also to cspF at a comparable region (the positioning of the primers is shown in Fig. 1). Furthermore, BcAPR1 hybridizes equally to both psychrotrophic and mesophilic B. cereus cspF sequences (both have identical complementary sequences). Hence, BcAPR1 in combination with BcFF2 can amplify a 284-bp cspF DNA fragment from both psychrotrophic and mesophilic strains of the B. cereus group whereas BcAPR1 in combination with BcAPF1 can only amplify a 160-bp cspA DNA fragment from psychrotrophic strains of this group (Fig. 1). Again, the PCR conditions used to accomplish the above amplifications were identical to those described previously, with the exception that the annealing temperature was 55°C (52°C worked equally well).

View larger version (45K): [in this window] [in a new window]   FIG. 1.   Schematic diagram of cspA and cspF, the two genes used as PCR target sequences for the differentiation of psychrotrophic and mesophilic strains of the B. cereus group (left). The approximate position at which each of the three primers hybridizes to each of the genes and the sizes of the resulting PCR products are indicated. Signature bases, allowing PCR amplification of psychrotrophic cspA only, are underlined and shown in grey. The right-hand side of the figure shows two pictures of an ethidium bromide-stained agarose gel, with examples of DNA amplification from psychrotrophic and mesophilic strains of the B. cereus group obtained with our PCR assay. The top picture shows amplified DNA products from 10 mesophilic strains (lanes: 2, WSBC10027; 3, WSBC10028; 4, HER1399; 5, ATCC 7064; 6, ATCC 27877; 7, HER1357; 8, CW1; 9, ATCC 14579T; 10, ATCC 6462T; 11, ATCC 10792T). The bottom picture shows amplified DNA products from 10 psychrotrophic strains (lanes: 13, WSBC10201; 14, WSBC10202; 15, WSBC10276; 16, WSBC10278; 17, WSBC10289; 18, WSBC10045; 19, WSBC10089; 20, CW9; 21, PG1; 22, ATCC 6462T). Lanes 1 and 12 were loaded with 100-bp DNA ladders. PCR was performed as described in the text, with an annealing temperature of 55°C.

In addition to testing the PCR assay described above with those strains employed in its construction (i.e., the strains with cspA signatures shown in Table 1), a large number of additional B. cereus, B. thuringiensis, and B. mycoides strains that were isolated from a wider variety of sources (in this study, unless otherwise stated) were also screened by the same PCR procedure (Table 1). Additional B. cereus strains were isolated from dairy products originating from nine other dairies situated throughout Germany (G3 to G11), as well as from dairy products from Denmark and Norway (P. Granum, Norwegian College of Veterinary Medicine, Oslo, Norway). Additional strains of B. thuringiensis were obtained from two other sources (E. Lonc, Wroclaw University, Wroclaw, Poland [10], and H. Ackermann, Reference Center for Bacterial Viruses, Quebec, Canada), while other B. mycoides strains were isolated from soil from Thailand. Type strains of B. cereus, B. thuringiensis, and B. mycoides were also included. Table 1 shows a summary of the data gained from these different strains, demonstrating in every case a correlation between the ability to grow at 7°C and the occurrence of a second 160-bp DNA fragment amplified by the above PCR assay.

To ensure that this PCR assay was specific for the detection of bacteria of the B. cereus group (i.e., B. cereus, B. thuringiensis, and B. mycoides) at an annealing temperature of 55°C (or 52°C), a large number of dairy-associated bacteria (both Bacillus and non-Bacillus strains) were also screened by the above-mentioned PCR procedure. The Bacillus species tested were B. brevis (ATCC 8246^T), B. circulans (ATCC 4513^T), B. coagulans (ATCC 7050^T), B. firmus (ATCC 14575^T), B. laterosporus (ATCC 64^T), B. lentus (ATCC 10840^T), B. licheniformis (ATCC 14580^T), B. macerans (ATCC 8244^T), B. megaterium (ATCC 14581^T), B. polymyxa (ATCC 842^T), B. pumilis (ATCC 7061^T), B. sphaericus (ATCC 14577^T), and B. subtilis (ATCC 6051^T). The non-Bacillus species tested were Alcaligenes faecalis (ATCC 8750^T), Arthrobacter nicotianae (ATCC 15236^T), Brevibacterium linens (WS 1978), Clostridium perfringens (ATCC 13124^T), Corynebacterium glutamicum (ATCC 14020), Enterobacter cloacae (ATCC 13047^T), Enterococcus faecalis (ATCC 19433^T), Escherichia coli (ATCC 11775^T), Flavobacterium flavescens (ATCC 8315), Klebsiella pneumoniae (ATCC 9997), Lactobacillus plantarum (ATCC 14917^T), Lactococcus lactis subsp. lactis (WS1042), Listeria monocytogenes (WSLC1364), Microbacterium lacticum (WS2024), Micrococcus luteus (WS1513), Proteus vulgaris (ATCC 13315^T), Pseudomonas fluorescens (DSM6147), Salmonella amersfoort (WS2686), and Staphylococcus aureus (ATCC 6538). No strains other than B. cereus, B. thuringiensis, and B. mycoides gave any amplification of PCR products. Moreover, 16S ribosomal DNA control PCRs, conducted in parallel, confirmed that PCR was capable of occurring efficiently for each bacterium (data not shown).

We believe that development of the above-mentioned PCR assay will significantly reduce the time necessary for screening pasteurized food and dairy products for contamination by psychrotrophic strains of the B. cereus group. It is also possible that employment of this assay during food processing will allow the number of such contaminating bacteria to be more accurately estimated, reducing the incidence of food poisoning by these potential pathogens.