- Top of page
- Materials and methods
Aims: To study the occurrence of Cronobacter spp. in foods and to investigate the phenotypic properties of the strains isolated.
Methods and Results: A total of 53 strains of Cronobacter spp. isolated from 399 food samples were identified using conventional biochemical methods and MALDI-TOF mass spectrometry. Foods of plant origin were the most frequently contaminated samples. No Cronobacter spp. were found in infant milk formula, wheat-based infant food, pasteurized and raw cow milk, mincemeat, chicken, chickpea and potato dumpling powder. The individual species were identified as Cronobacter sakazakii (54·7%), Cronobacter malonaticus (28·4%), Cronobacter dublinensis (7·5%), Cronobacter muytjensii (7·5%) and Cronobacter turicensis (1·9%). Cronobacter sakazakii and C. malonaticus belong to biotype 1, 2, 2a, 3, 4 and 5, 5a, respectively. Cronobacter dublinensis strains were subdivided into biotypes 6 and 12. All strains were resistant to erythromycin and two of them were resistant to both erythromycin and tetracycline.
Conclusions: Cronobacter spp. were isolated from various food samples pre-eminently of plant origin and dried food ingredients.
Significance and Impact of the Study: These findings will increase and detail our knowledge of the presence and diversity of Cronobacter spp. in foods.
- Top of page
- Materials and methods
Cronobacter sakazakii and related species are opportunistic pathogens causing rare but life-threatening diseases such as meningitis, necrotizing enterocolitis and bacteremia in premature, low-birth and immunocompromised infants (FAO/WHO 2008). Neonates and infants with Cronobacter meningitis have poor prognosis. The lethality rate was estimated to be 41·9% (Friedemann 2009), with death occurring within hours after manifestation of first symptoms (Willis and Robinson 1988). These bacteria have been also recognized as the causative agents of various infections in elderly adults suffering from serious underlying disease or malignancy (Dennison and Morris 2002; Gosney 2008).
Cronobacter spp. are Gram-negative, facultative-anaerobic, nonspore-forming, motile bacteria belonging to the family Enterobacteriaceae. The genus comprises six species: C. sakazakii, Cronobacter dublinensis, Cronobacter malonaticus, Cronobacter muytjensii, Cronobacter turicensis and the unnamed genomospecies I (Iversen et al. 2006, 2008). The natural habitat of these bacteria is unknown. However, some features such as the production of yellow pigment, siderophores and indole acetic acid, their ability to persist in a desiccated state, solubilize mineral phosphate and to colonize tomato and maize roots, all indicate that Cronobacter species are probably plant-associated micro-organisms (Schmid et al. 2009). Owing to their ubiquitous nature, Cronobacter spp. have been isolated from various environmental samples such as soil (Neelam et al. 1987), water (Friedemann 2007), rats (Gakuya et al. 2001), insects (Kuzina et al. 2001; Hamilton et al. 2003; Butler et al. 2010), dust from household and milk powder factories (Kandhai et al. 2004), and domestic brushes, stirring spoons and blenders used to mix infant formula (Muytjens et al. 1983; Block et al. 2002). Moreover, Cronobacter (Enterobacter) sakazakii has been isolated from a wide spectrum of foods and food ingredients, as reviewed by Friedemann (2007). It was recovered from plant food including vegetables, herbs, spices, legumes and cereal products, as well as foods of animal origin such as milk, cheese, fish, meat and meat products. The bacteria were also found in fresh, dry, frozen, cooked and ready-to-eat products. Although Cronobacter has been found in numerous types of foods, only infant milk formulae and follow-up formulae have been associated with infant infections (Jarvis 2005). Raw materials used for preparation of these foods may be important sources of pathogenic micro-organisms, but these bacteria including Cronobacter species are eliminated during the pasteurization. The contamination with Cronobacter occurs during the addition of dry ingredients and fortifiers, as well as during postpasteurization handling and packaging (Iversen and Forsythe 2003; Mullane et al. 2007). Hygienic procedures applied in the production facility also have an impact on the ability of Cronobacter spp. to persist and proliferate in the manufacturing environment and to contaminate the final product.
There is only fragmentary knowledge of the presence of Cronobacter spp. in foods commercially available in the Czech Republic. Therefore, this study aimed to investigate the occurrence of these micro-organisms in a wide spectrum of retail foods. Various food products including infant milk formula, wheat-based infant food, milk, milk powder, spices and herbs, legumes and meat were screened, and the bacteria isolated were identified by means of standard biochemical tests and matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF). All strains were biotyped and their antimicrobial susceptibility was established.
Materials and methods
- Top of page
- Materials and methods
Nutrient Broth and Peptone Water were purchased from Oxoid (Basingstoke, UK); Difco MIO Medium was obtained from Bencton, Dickinson & Co. (Sparks, MD, USA); Methyl Red Voges–Proskauer Broth, Methyl Red Indicator, Malonate Broth, Nitrate Broth, Kovác’s reagent for indoles and dulcitol were purchased from Sigma-Aldrich (Prague, Czech Republic); Nutrient Agar, Tryptic Soy Agar, Lauryl Sulfate Broth, Griess-Illosvay’s reagent, zinc dust and vancomycin hydrochloride from Streptomyces orientalis were purchased from Merck, KGaA (Darmstadt, Germany); Peptone Water was purchased from Oxoid CZ (Brno, Czech Republic); Peptone Water with Phenol Red, Mueller-Hinton Agar and discs containing antibiotics were obtained from HiMedia (Mumbai, India); ESIA Enterobacter Sakazakii Isolation Agar (AES Chemunex, Bruz, France) was obtained from O.K. Servis Bio Pro (Prague, Czech Republic); ENTEROtest24 kit and OXItest strips were purchased from Pliva-Lachema (Brno, Czech Republic); API ID 32E strips were purchased from bioMérieux CZ (Prague, Czech Republic); CCA matrix and Bacterial Standard Test were obtained from Bruker Daltonik GmbH (Bremen, Germany).
A total of 399 samples of 34 types of foods and ingredients were tested during the period from January to December 2010 for the presence of Cronobacter spp. With the exception of raw milk, foods were purchased from retailers located within Prague city. Raw milk was obtained from vending machines located in Prague (n = 8), Liberec (n = 3) and Soběslav (n = 5) and from a farm at Jeřmanice (n = 6). The food samples of animal and plant origin included powdered infant milk formula intended for infants above 4 months old, wheat-based infant food intended for infants above 6 months old, pasteurized and powdered milk, cereals, legumes, spices, herbs, chicken, mincemeat, eggs, instant lentil soup, desiccated coconut, potato dumpling powder, tea and various seeds (Table 1).
Table 1. Occurrence of Cronobacter spp. in retail foods
|Samples||Number of samples||Positive samples||Species|
| Infant milk formula||8||0|| || || || || |
| Wheat-based infant food||10||0|| || || || || |
| Pasteurized milk||10||0|| || || || || |
| Powdered milk||20||2||2|| || || || |
| Powdered goat milk drink||10||0|| || || || || |
| Raw milk||22||0|| || || || || |
|Eggs and meat|
| Eggs||10||1||1|| || || || |
| Chicken||35||0|| || || || || |
| Mincemeat||10||0|| || || || || |
| Doughnut||10||0|| || || || || |
| Pasta||8||0|| || || || || |
| Oat flakes||10||1||1|| || || || |
| Rice flour||16||6||1||3|| ||2|| |
| Wheat sprout||9||1||1|| || || || |
| Chickpea||8||0|| || || || || |
| Instant lentil soup||10||6||2||4|| || || |
| Lentils||11||1|| || || ||1|| |
| Tofu||11||4||3|| || ||1|| |
|Spices and herbs|
| Basil||10||4||1|| ||3|| || |
| Caraway||15||2||2|| || || || |
| Cinnamon||11||0|| || || || || |
| Ginger||10||3||2||1|| || || |
| Marjoram||12||2||1|| ||1|| || |
| Oregano||10||0|| || || || || |
| Pepper||14||1||2|| || || || |
| Pimiento||10||2|| ||2|| || || |
| Sweet pepper||12||0|| || || || || |
| Poppy seed||11||1||1|| || || || |
| Pumpkin seed||11||5||2||2|| || ||1|
| Sesame seed||12||8||6||2|| || || |
| Desiccated coconut||10||1||1|| || || || |
| Coconut biscuits||1||1|| ||1|| || || |
| Potato dumpling powder||14||0|| || || || || |
| Tea||8||0|| || || || || |
Enrichment and isolation of Cronobacter spp.
The enrichment and isolation of Cronobacter spp. were performed according to procedures described in the directive ČSN P ISO/TS 22964 (2006) Milk and milk products – Detection of Enterobacter sakazakii. Aliquots (25 g) of each of the studied samples were weighed separately and aseptically into a sterile Stomacher bag, diluted with 225 ml of Peptone Water and homogenized in a peristaltic homogenizer for 2 min. If the content of a food package weighed <25 g, the quantities were proportionally smaller. The homogenate was incubated at 37°C for 18 h. Then 0·1 ml of the pre-enriched sample was transferred into 10 ml of Lauryl Sulfate Broth supplemented with NaCl to a final concentration of 34 mg ml−1 and vancomycin to a final concentration of 0·01 mg ml−1. The sample was incubated at 44°C for 24 h. One loop of the selectively enriched broth was streaked on the surface of ESIA Enterobacter Sakazakii Isolation Agar and incubated at 44°C for 24 h. Green or blue–green colonies were termed presumptive Cronobacter spp. and their identity was confirmed by means of biochemical tests and MALDI-TOF mass spectrometry.
Identification and biotyping of Cronobacter spp. by biochemical tests
In addition to Gram staining and phenotype characterization derived from commercial biochemical ENTEROtest24 kit and API ID 32E strips according to producers’ instructions, the following tests were performed using conventional manual methods. Motility was determined at 37°C after 24 and 48 h using Difco MIO medium. The methyl red test was performed by adding 0·5 ml of indicator to a culture grown in 5 ml of Methyl Red Voges–Proskauer Broth in a 16 by 125 mm tube at 37°C for 48 h. Reduction of nitrate to nitrite was tested by addition of Griess–Illosvay’s reagent to cultures grown under anaerobic conditions in 5 ml of Nitrate Broth for 48 h. Zinc dust was added to negative tubes of nitrate negative strains to confirm the presence of unreduced nitrate. Gas production from d-glucose was tested in 5 ml of Peptone Water containing 0·5% (w/v) of d-glucose. An inoculated medium was incubated at 37°C for 24 h. Developing gas was collected in Durham tube inserts (Durham 1898). Yellow pigment production was observed for cultures grown on Tryptic Soy Agar incubated at 25°C for 2–5 days. Production of catalase was determined by dropping 0·05 ml of 3% hydrogen peroxide on 24 h old Nutrient Agar cultures. The cytochrome oxidase test was performed with commercial OXItest strips according to the producer’s instructions.
The following phenotypic attributes were verified by conventional tube tests: indole production, utilization of malonate and production of acid from dulcitol. Indole production was measured by the addition 0·5 ml of Kovác’s reagent to cultures grown in Peptone Water at 37°C for 48 h. Malonate utilization was tested by incubation in Malonate Broth at 37°C for 24–48 h. Acid production from dulcitol was tested in 5 ml of Peptone Water with Phenol Red, containing dulcitol (final concentration 0·5%) added before the autoclaving step.
Identification of Cronobacter spp. by MALDI-TOF spectrometry
Bacterial strains tested were grown on Tryptic Soy Agar at 28°C for 24 h. A small amount of fresh culture was transferred directly onto the ground-steel MALDI target using a sterile tip and dried at room temperature. Subsequently, the dried cell spots were overlaid with CCA matrix solution (50% acetonitrile-2·5% trifluoroacetic acid). The Bruker Bacterial Standard Test was mixed with CCA matrix solution in the ratio of 1:2 (v/v) and spotted onto the ground-steel MALDI target as well. After crystallization, the samples were analysed using a Biflex IV mass spectrometer (Bruker Daltonik) equipped with a UV nitrogen laser (337·1 nm) at 9 Hz repetition rate (Laser Science, Inc., Franklin, MA, USA) in linear mode (for the mass range 2–20 kDa) using an accelerating voltage of 19 kV, a lens voltage of 7·75 kV and 200 ns of delay time. The raw spectra of unknown micro-organisms were imported into the MALDI BioTyper™ software and processed and analysed with standard pattern matching against reference spectra of 3287 micro-organisms in the MALDI BioTyper™ reference database (Bruker Daltonik).
The resistance profile of strains isolated was determined using the standardized Bauer–Kirkby agar disc diffusion method (Bauer et al. 1966). Each isolate was inoculated on Mueller–Hinton agar according to the WHO laboratory protocol (WHO 2010). Antimicrobial-containing discs that included ampicilin 10 μg, piperacilin 100 μg, ceftriaxone 30 μg, cefuroxime 30 μg, imipenem 10 μg, amikacin 30 μg, gentamicin 10 μg, erythromycin 15 μg, tetracycline 30 μg, ciprofloxacin 5 μg, nalidixic acid 30 μg, norfloxacin 10 μg, chloramphenicol 30 μg, nitrofurantoin 300 μg, trimethoprim 5 μg or co-trimoxazole 25 μg were applied to the surface of the inoculated plates and incubated at 37°C for 24 h. The diameters of the inhibition zones were measured and interpreted according to a table supplied by the manufacturer.
- Top of page
- Materials and methods
Foods of plant origin were the most frequently contaminated samples. The highest percentage of Cronobacter spp. relative to the total number of isolates was found in herbs and spices (28·3%), seeds (26·4%) and cereals (15·1%). Powdered milk samples were rarely contaminated with target micro-organisms. No Cronobacter spp. were isolated from infant milk formula, wheat-based infant food, pasteurized and raw cow milk, mincemeat, chicken, chickpea, pasta and potato dumpling powder. These results correspond to those of previous studies that focussed on isolation of Cronobacter from foods. Iversen and Forsythe (2004) isolated Cronobacter spp. (E. sakazakii) from 44 of 122 samples of herbs and spices and 15 of 66 other dried food ingredients. But only three of 72 powdered milk samples were positive for the presence of these bacteria. Baumgartner et al. (2009) isolated these micro-organisms from confectionery, sprouts, fresh herbs, fresh salads, spices and dried herbs. On the other hand, negative results were obtained in soft cheeses from raw milk, meat products, delicatessen salads, ice cream and milk and dessert powders. Jaradat et al. (2009) reported the highest presence of Cronobacter spp. in herbs and spices among tested foods but no positive results in powdered milk. Similarly, Kandhai et al. (2010) found Cronobacter spp. in dry cereals, beef, mixed beef and pork, dry vegetables and spices but not in raw cow milk samples. According to Baumgartner and Niederhauser (2010), no Cronobacter spp. were present in 875 raw milk samples, and milk appeared not to be the source of these micro-organisms with the contamination of powdered milk and other milk products originating instead from the environment.
The occurrence of cronobacters in the studies of infant milk formulae has declined in the last two decades from 14·2% (Muytjens et al. 1988) to c. 1–2% (Iversen and Forsythe 2004; Shaker et al. 2007; Chap et al. 2009; O’Brien et al. 2009. The absence of Cronobacter spp. in our samples of infant milk formula intended for infants above 4 months old and wheat-based infant food intended for infants above 6 months old was probably not only due to the application of good manufacturing practice by producers but also due to the limited number of tested samples.
The isolates were identified by means of biochemical testing of 48 phenotypic attributes and using MALDI-TOF mass spectrometry. The latter analytical system has already been successfully used in several studies for the identification of food-borne pathogens such as Campylobacter, Listeria and Salmonella (Barbuddhe et al. 2008; Dieckmann et al. 2008; Alispahic et al. 2010). Although the mass spectra of the isolates were compared with those of 3287 micro-organisms, our analytical system enabled us to identify target bacteria only to Cronobacter genus level. Despite these limitations, the MALDI-TOF mass spectrometry proved to be a reliable, rapid and powerful technique for the identification of the micro-organisms in question, especially in combination with biochemical tests. This is particularly true with the advent of genus- and species-identifying biomarker mass patterns for the detailed identification of Cronobacter spp., which have been described recently (Stephan et al. 2010).
As mentioned above, almost all known species of Cronobacter were isolated from our food samples. Our findings are partially in agreement with the results of Molloy et al. (2009). These authors found C. sakazakii (21 isolations), C. malonaticus (11 isolations) and C. turicensis (1 isolation) in 518 various food and environmental samples. The identification was based on biochemical tests for indole production, malonate utilization and dulcitol fermentation. The only significant species C. sakazakii and C. malonaticus were reported by Mullane et al. (2008) focussed on the spreading of Cronobacter in the production of a powdered milk protein manufacture.
It is likely to be significant both ecologically and clinically that the bacteria recovered from our food samples belong to only a few biotypes. Out of a possible 16 Iversen’s et al. (2006) biotypes and subtypes of C. sakazakii, only biotypes 1, 2, 2a, 3 and 4 were found, with biotypes 5 and 5a of C. malonaticus. Infrequent C. dublinensis strains, isolated from lentils, tofu and rice flour, were divided into C. dublinensis subsp. dublinensis and C. dublinensis subsp. lactaridi. Cronobacter spp. isolated from various sources are usually genotyped, but detailed information about the biochemical properties of the isolates is missing from the most published reports. Such information could be helpful in studies focussing on the virulence of clinical and environmental strains or those related to the ecology of these micro-organisms.
The 53 Cronobacter isolates obtained in this study were resistant to erythromycin, but varied in their susceptibility to tetracycline, with 35·8% of strains being susceptible to tetracycline, 60·4% of strains showing moderate resistance and 3·8% of micro-organisms being resistant to this antibiotic. All micro-organisms isolated from food samples were susceptible to β-lactams, cephems, carbapenems, aminoglycosides, quinolones, nitrofuran compounds, chloramphenicol, diaminopyrimidine derivative trimethoprim and co-trimoxazole. These results correspond to the findings of Stock and Wiedemann (2002) for Enterobacter (Cronobacter) sakazakii. Terragno et al. (2009) tested the susceptibility of 23 Cronobacter spp., isolated from powdered infant formulae, to amoxicillin/clavulanic acid, ampicilin, cefotaxime, ceftriaxone, chloramphenicol, ciprofloxacin, gentamicin, nalidixic acid, tetracycline, co-trimoxazole, cefuroxime and cefixime. These authors found that all of their isolated strains were susceptible to all the antibiotics tested. Kim et al. (2008) described Cronobacter spp. strains recovered from various foods that were susceptible to tetracycline and resistant to ampicillin or cephalothin. Cronobacter spp. isolated from Mexican fruit flies Anastrepha ludens were also resistant to ampicillin and cephalothin, as well as to erythromycin, novobiocin and penicillin, but they were susceptible to tetracycline (Kuzina et al. 2001). In contrast, Nazarowec-White and Farber (1999) found that two of eight Cronobacter strains recovered from food samples showed resistance to chloramphenicol and tetracycline. Farmer et al. (1980) described the susceptibility pattern of 24 E. sakazakii (Cronobacter spp.); all strains were susceptible to gentamycin, kanamycin and chloramphenicol, 97% were susceptible to nalidixic acid, 92% were susceptible to streptomycin, 87% were susceptible to tetracycline and carbenicillin, 67% were susceptible to sulfadiazine and 13% were susceptible to cephalothin. All strains were resistant to penicillin. These authors noted that only 1 strain out of >100 Cronobacter strains tested were multiresistant to streptomycin, kanamycin, tetracycline and chloramphenicol.
The members of the genus Cronobacter are ubiquitous micro-organisms that have been found in various foods particularly of plant origin and dried material. The bacteria are able to survive desiccation and thus persist in viable form for a long period. On the one hand, their low prevalence in products of animal origin indicates that these types of samples do not represent inherent sources of Cronobacter spp. These foods are inevitably contaminated during some stages of the manufacturing process. Numerous foods and ingredients tested in the present study are ones that are normally heated before consumption. Cronobacter spp. did not usually survive under such conditions, but there is still the risk of accidental cross-contamination in household kitchens and restaurants. An impact of the consumption of contaminated foods on elderly or immunocompromised individuals is not sufficiently known. In comparison with some previous reports, no Cronobacter spp. were detected in powdered infant milk formula and wheat-based infant food. This phenomenon probably resulted from the application of good manufacturing practice and respecting the strict hygiene procedures during the manufacturing process. On the other hand, it must be emphasized that only limited numbers of these samples were tested in this study.
MALDI-TOF mass spectrometry was proven to be a reliable and rapid tool for Cronobacter detection if it is used simultaneously with biochemical testing. An application of at least two identification methods is necessary to eliminate misidentified strains. Should one of these techniques fail and give discrepant results, one can use additional methods such as PCR analysis of specific genes, fatty acid profile analysis or 16S rRNA sequencing for the reliable identification of Cronobacter spp.