Cronobacter Species in Powdered Infant Formula and Their Detection Methods

Cronobacter species are opportunistic pathogens, which cannot cause infectious disease when ingested by healthy adult but pose a very high risk to neonates and immuno-compromised individuals (Healy et al., 2010; Strydom et al., 2012). These organisms can affect the human central nervous system, and survivors often suffer from severe neurological impairments following infection, including hydrocephalus, quadriplegia, and developmental delays (Gallagher and Ball, 1991; Lai, 2001; Gurther et al., 2007; Strydom et al., 2012). Cronobacter species have caused diseases in all age groups, most importantly in infants, and pre-term, underweight, or immune-compromised infants at particularly high risk (FAO/WHO, 2004).

Cronobacter species have been isolated not only from plant food and products such as cereals, fruits and vegetables, legume, herbs, and spices, but also from the animal foods such as milk, meat, fish, and products made from these foods (Lee et al., 2012). Despite the reservoir and mode of transmission of Cronobacter species remain unknown, they are believed to be the principal sources of food contamination. Soil, water and vegetables, whereas rodents and flies may serve as a secondary route of contamination (Iversen and Forsythe, 2003). PIF is the only food source that has been epidemiologically linked to outbreaks of Cronobacter species (FAO/WHO, 2004).

There are many reported cases of Cronobacter species since the first two documented cases of infection occurred in 1958 at St. Albans, England (Urmenyi and Franklin, 1961). Outbreaks and cases of Cronobacter species until 2005 had been well summarized by many researchers as well as showen in Table 1 (Gurtler et al., 2007; Bowen and Barden, 2006; Iversen and Forsythe, 2003). Pagotto and Farber (2009) reported a 17 days infant with born weight of 2.9 kg developed a Cronobacter infection leading to meningitis after fed infant formula powder that had been left at her doorstep. Centers for Disease Control and Prevention (CDC) (2009) reported 2 unrelated cases of infection of Cronobacter species in New Mexico, USA. A female infant was born in September 2008 at full term by vaginal delivery without complications and had been fed PIF exclusively since birth. Cronobacter species was positive in Cerebrospinal fluid cultures but was negative in blood cultures. After treatment in the hospital, the infant was discharged on February 6, 2009, with severe brain injury, hypertonicity resulting from central nervous system damage, and hydrocephalus. A male infant was born in April 2008 at 40 weeks' gestation by vaginal delivery without complications and breastfed exclusively until age 6 months and then began transitioning to PIF and age-appropriate foods. A Cronobacter organism was found in the blood culture obtained during autopsy. The two infants lived approximately 200 miles apart in different counties of mostly rural southeastern New Mexico. Flores et al. (2011) reported two cases of acute gastroenteritis occurred in 5-month-old infants hospitalized in a mother-and-child hospital. C. sakazakii was recovered from the PIF, rehydrated infant formula powder fed to infants, and their fecal samples and no deaths were observed in those cases. CDC (2012) reported 4 infants (children under 12 months of age) infected with Cronobacter bacteria resided in four states: Florida, Illinois, Missouri, and Oklahoma. One infant was born in slightly premature, and all four infants became infected during the first five weeks of age. Three infants developed meningitis (inflammation of the membranes around the brain and spine), and one had a bloodstream infection and 2 infants died. Broge and Lee (2013) reported a female infant born at 37 weeks via vaginal delivery to a 22-year-old mother. On day of life 33, the infant began sleeping more than usual and was not feeding well. Her mother recorded an axillary temperature of 38.3°C, prompting a visit to her pediatrician. Cronobacter species was positive from blood cultures. The infant was discharged after treatment. Recently, Cui et al. (2017) reported two cases of Cronobacter species infection of infants in China. The first case was of a female infant born after 36 weeks of gestation that developed fever and muscle spasms in her extremities and fed poorly after 11 days of birth. Cronobacter was isolated from the cerebral spinal fluid using selective medium. The second case was an abandoned female infant with no detailed birth record. She was fed with PIF and developed anemia, diarrhea, moderate dehydration, and severe malnutrition after an operation when she was 10 days old. The infant died due to circulatory and respiratory failure after 4 days in hospital. Cronobacter was isolated and identified from the blood sample of infected infant.

Cronobacter species can be isolated from a range of environments, as well as clinical, food, and beverage sources such as water, vegetables, cheese, meat and ready-to-eat foods. Most of the documented outbreaks of Cronobacter species have been associated with consumption of contaminated PIF. Therefore, isolations of Cronobacter species from infant food and dairy food in recent years are summarized in Table 2. According to Jung and Park (2006), 20% of infant formula powder was contaminated with Cronobacter in the Republic of Korea. A study carried out by Shaker et al. (2007) showed that Cronobacter species have been isolated from 17% of PIF, all the food samples were purchased from retail stores across Jordan. Fu et al. (2011) also reported that 4% (3/77) of analyzed PIF were positive for the Cronobacter species. Lee et al. (2012) also reported that Cronobacter species were isolated from 18.6% of detected food samples. Li et al. (2014) analyzed 195 food samples, including cereals, cereal products, PIF, infant food formula, herbs, spices, vegetables, and fruits collected from the local supermarkets in Nanjing, China for the presence of Cronobacter species. PIF and infant food formula were negative while 14.1% (12/85) cereal and cereal products and 4.5% (1/22) spices were positive for Cronobacter species. Pei et al. (2016) collected 2,282 samples of PIF and follow-up formula from the retail markets in China from January, 2012 to December, 2012 to determine Cronobacter species contamination in PIF and follow-up formula in China. Cronobacter species have been isolated from 1.1% (25/2,282) collected samples, and the positive rates in infant formula powder and follow-up formula were 0.90% (10/1,011) and 1.18% (15/1,271). Pan et al. (2014) also surveyed 399 samples of PIF and follow-up formula purchased from retail stores and supermarkets in Shijiazhuang, China. The positive rates from PIF and follow-up formula were 11.5% (19/165) and 12.8% (30/234), respectively, including 48 C. sakazakii and 1 C. malonaticus. These researchers reported the analysis and isolates of Cronobacter species in PIF, despite the source of contamination have not mentioned. Yao et al. (2016) isolated Cronobacer species from 12% (16/133) of indigenous infant flours sold in public heath care centers with in Abidjan, Cote d’lvorie.

Culture based detection is traditional and most widely used method for detecting foodborne pathogens. The culture methods most often used for the detection of Cronobacter species are based on the United States Food and Drug Administration (USFDA) method for isolation and enumeration of PIF (Druggan and Iversen, 2009; Strydom et al., 2012). In 2002, the USFDA devised a most probable number method for isolation and identification of Cronobacter species in PIF. However, this method requires food samples to be pre-enriched, and 5-7 days are required to identify organisms and confirm the results (USFDA, 2002). For this method, 1 g, 10 g, and 100 g of infant formula powder are rehydrated in sterile distilled water and incubated at 36°C overnight to resuscitate stressed cells, after which 10 mL of sample are taken from the pre-enrichment broth and added into 90 mL of Enterobacteriaceae Enrichment broth for incubating at 36°C overnight. Next, the enrichment broth is spread or streaked onto violet red bile glucose agar (VRBGA), after which the colonies selected by VRBGA are streaked onto tryptone soya agar (TSA). Following cultivation, samples are tested using an API 20E kit for identification.

In 2012, the USFDA released a new method for isolation and enumeration of Cronobacter from formula samples and confirmation of Cronobacter using PCR screening and selective media (USFDA, 2012). In this method, formula samples are incubated in buffered peptone water (BPW) at a 1:10 ratio and 36°C for 24±2 h without shaking. At 6-8 h, four aliquots of 40 mL enrichment culture of formula sample are removed from the culture and centrifuged at 3,000×g for 10 min to collect the Cronobacter. The collected bacterial cells are then suspended in 200 μL of phosphate buffered saline. Two of these samples are used for real-time PCR and the other two are plated onto two Druggan-Forsythe-Iversen (DFI) chromogenic agar plates and two R & F E. sakazakii chromogenic agar plates. If the 6-8 h culture are confirmed as positive by real-time PCR and RAPID ID 32E, no further analysis is necessary, otherwise, the 18-24 h culture is processed and analyzed (USFDA, 2012).

The International Organization for Standardization (ISO, 2006) also released a method for the detection of Cronobacter species in milk and milk products. In this method, sample was incubated in BPW at a 1:10 proportion at 37°C for 18±1 h without shaking (ISO, 2006). Next, 0.1 mL of culture was inoculated into 10 mL of Modified Laurylsulfate-Tryptose broth with vancomycin and incubated for 24±2 h at 44±0.5°C, followed by streaking onto selective chromogenic agar and incubating for 24±2 h at 44±1°C. Five colonies selected from selective chromogenic agar were then streaked onto TSA plates and incubated for 24±2 h at 25±1°C, after which a colony from each TSA plate was tested by biochemical identification (ISO, 2006). The ISO method also requires 5 days to complete all tests.

In addition to the FDA and ISO methods, several selective media have been developed for the isolation and detection of Cronobacter species. Iversen et al. (2008) developed a differential medium, Cronobacter screening broth (CSB), to complement agars based on hydrolysis of chromogenic α-glucopyranoside substrates. In this method, spiked samples are incubated for 24 h at 37°C, then diluted to 1:10 in CSB and incubated for 24 h at 42°C. Each sample is then streaked onto TSA that is subsequently incubated at 37°C for 24 h to detect presumptively positive Cronobacter colonies. The presumptively positive Cronobacter colonies are distinguished using the methyl red and Voges-Porskauer tests. Kim and Rhee (2011) also reported a selective and differential medium for the isolation of Cronobacter species, which contained salicin as a differential agent. In this new medium, Cronobacter species generates typical colonies with characteristic violet-colored centers surrounded by a transparent to opalescent border, and the growth of other microorganisms are inhibited or produce visually distinguishable colonies. Oh and Kang (2004) developed a fluorogenic selective and differential medium for isolation of Cronobacter species using 4-methylumbelliferyl-D-glucoside as a selective marker. For this technique, C. sakazakii mixed with cocktail is plated onto OK medium and incubated, resulting in distinct fluorescent colonies when exposed to long-wavelength UV light. The fluorescent colonies are then confirmed using the API 20E biochemical system.

PCR is the most well-known and established nucleic acid amplification technique for detection of pathogenic microorganisms (DeCory et al., 2005; Stoop et al., 2009; Zhao et al., 2013). In this method, double-stranded DNA is denatured into single strands, and specific primers or single-stranded oligonucleotides anneal to these DNA strands, followed by extension of the primers complementary to the single stranded DNA with a thermostable DNA polymerase. These steps are repeated, resulting in doubling of the initial number of target sequences with each cycle. The quantity of the products of amplification can be visualized as a band on an ethidium-bromide-stained electrophoresis gel. Identification based on PCR amplification of target genes by sequencing is considered to be a reliable technique when properly developed and validated for a certain species. With the distinct advantages of rapidity, specificity, sensitivity, and fewer samples over culture-based methods, many PCR assays for the detection and validation of foodborne bacteria and viruses in food have been developed and applied. In addition to PCR, a number of gene-specific hybridization probes have been designed and used for the detection of toxin genes in foodborne pathogens (DeCory et al., 2005; Ikeda et al., 2007; Xu et al., 2014a; Zhao et al., 2014). A new method combining a real-time PCR based assay with selective agar has been also developed that can efficiently reduce the detection time to 2 days (Lampel and Chen, 2009). Stoop et al. (2009) reported a different rpoB based PCR system that enabled identification to the species level of strains previously confirmed to belong to the genus Cronobacter by using designed primer pairs based on the rpoB sequences of six Cronobacter species type stains. Zimmermann et al. (2014) developed a fast and sensitive PCR based detection system for Cronobacter species that consisted of enrichment, DNA-isolation and detection by real-time PCR using the outer membrane protein gene ompA as a target. Zhou et al. (2008) reported an immobilization and detection method of C. sakazakii from PIF. In this method, C. sakazakii cells were immobilized with zirconium hydroxide and separated from the sample by centrifugation. The immobilized cells were then suspended in sterile nutrient broth and subjected to DNA extraction for duplex PCR detection. Ruan et al. (2013) reported a method of detecting Cronobacter species by duplex PCR in combination with a capillary electrophoresis-laser induced fluorescence detector. The specific gene sequences of the 16S-23S rDNA internal transcribed spacer and the outer membrane protein A (OmpA) of Cronobacter species were amplified by duplex PCR. The PCR products were separated and determined sensitively by capillary electrophoresis-laser-induced fluorescence within 12 min. Wang et al. (2012) reported a TaqMan real-time PCR assay incorporating an internal amplification control for the detection of Cronobacter species. In this study, the sensitive and accurate TaqMan real-time PCR assay was improved by the introduction of internal amplification control to control false-negative results that may be caused by malfunction of the thermal cycler, incorrect PCR mixture, inhibitory substances, etc.

Methods based on antigen-antibody binding are widely used to identify foodborne pathogens. An indirect non-competitive enzyme-linked immunosorbent assay (INC-ELISA) (Song and Kim, 2013) and a sandwich ELISA (Park et al., 2012) were reported for the detection C. muytjensii in formula powder. These methods were sensitive and efficient for the detection of C. muytjensii using anti-C. muytjensii IgG. Similarly, Xu et al. (2014b) reported two novel analytical methods, sandwich ELISA and indirect enzyme-linked ELISA, based on polyclonal and monoclonal antibodies for rapid detection of Cronobacter sakazakii. Song et al. (2015) reported a fluorescence-based liposome immunoassay for detection of C. muytjensii that used developed anti-C. muytjensii IgG targeted liposome. In this study, samples or cultures were coated onto 96 well plates and anti-C. muytjensii IgG targeted liposome was added and incubated. The immunoliposome bounded to the cell was lysed and measured at an excitation wavelength of 550 nm and an emission wavelength of 585 nm using the microplate reader. Blazkova et al. (2011) reported an immunochromatographic strip test for detection of members of the genus Cronobacter. Following the shortened bacteria cultivation and isolation of DNA, a specific gene sequence targeting 16S rRNA from Cronobacter species was amplified by PCR using 5-end labeled specific primers. The PCR product, amplicon labeled with digoxigenin on one side and biotin on the other side, was added directly to a immunochromatographic strip test composed of nitrocellulose membrane with bound antibody against digoxigenin in the test line. The visualization was mediated by colloidal carbon conjugated to neutravidin, and the appearance of a grey/black line was indicative of the presence of a specific amplicon. The color intensity of the test line in the pathogen-positive assay was visually distinguishable from that of negative samples within 10 min. Zhao et al. (2010) reported a method of cross-priming amplification under isothermal conditions combined with immuno-blotting analysis. Briefly, a set of specific displacement primers, cross primers and testing primers were designed based on six specific sequences in the Cronobacter species 16S-23S rRNA internal transcribed spacer. The specific amplification and hybridization steps were processed simultaneously under isothermal conditions at 63°C for 60 min.

Tools for on-site detection of Cronobacter species are also receiving a great deal of attention in recent years (Vigayalakshmi, 2010). Several rapid methods to detect Cronobacter species have been reported. Chen et al. (2014) reported an immunochromatographic strip for rapid detection of Cronobacter species in powdered infant formula in combination with silica-coated magnetic nanoparticles separation and a 16S rRNA probe. In this method, silica-coated magnetic nanoparticles were used to separate nucleic acid from Cronobacter cell lysate and eliminate the interference of food matrices and two 5’-end labeled probes that were complementary to the 16S ribosomal DNA of Cronobacter were used to hybridize with the nucleic acid. The hybrid product, labeled with digoxigenin on one side and biotin on the other side, was submitted directly to the immunochromatographic strip test which anti-digoxigenin monoclonal antibody was immobilized on nitrocellulose membrane in the test line. Visualization was achieved using gold nanoparticles conjugated to streptavidin, and double red bands appearing in both the test and control line indicated the presence of Cronobacter in the sample. Xu et al. (2014a) reported a method of probe-magnetic separation PCR for detection of Cronobacter species. In this method, a specific probe for Cronobacter species was designed, verified and conjugated on magnetic-beads. The targeted gene captured by the probe-magneticbead was separated from tested samples by magnetic force. Next, the target gene captured by probe-magneticbeads, and the total number of target sequences in each dilution was detected with primer pair CsspF₁ and CsspR₁ by using real-time qPCR. This system is able to detect as few as 10³ CFU/mL of Cronobacter species in artificially contaminated PIF in less than 4 h, without prior lengthy preenrichment culturing of samples. Shukla et al. (2016) developed an immunoliposome-based immunomagnetic concentration and separation assay for the rapid detection of C. sakazakii. In their method, C. sakazakii cells were separated and concentrated by magnetic nanoparticles (diameter 30 nm) which coated with anti-C. sakazakii IgG, followed by reaction with immonoliposomes which were modified with anti-C. sakazakii IgG. After removing the free immunoliposomes by washing, the immunomagnetic nanoparticle-C. sakazakii-immunoliposme complexes was mixed with n-octyl-β-D-glucopyranoside solution and measured an excitation wavelength of 550 nm and an emission wavelength of 585 nm using a microplate reader. The detection limit of this method was around 10³ CFU/mL in both pure culture and in PIF. Mullane et al. (2006) developed a method using cationic-magnetic-beads to capture C. sakazakii, followed by identification after plating the captured C. sakazakii onto DFI chromogenic agar, which can detect 1-5 CFU/500 g of PIF in less than 24 h. Zhao et al. (2013) reported an immunoassay based on nuclear magnetic resonance for detection of Cronobacter species in dairy samples with biofunctionalized magnetic nanoparticles. In this method, the amino-modified silica-coated Fe₃O₄ was conjugated with anti-C. sakazakii-antibody. C. sakazakii bound to the antibody-modified nanoparticles could be separated and measured based on their nuclear magnetic resonance using magnetic resonance relaxometry.

Mass spectrometry based on intact cell matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) protein profiling is a rapid and reliable tool for identification of microorganisms that is now widely used, even in clinical laboratories (Koubek et al., 2012; Krasny et al., 2014). Javurkova et al. (2012) tested MALDI-TOF mass spectrometry for the detection of Cronobacter species and compared the results with those obtained by the classical cultivation microbiological method. Javurkova’s method is specific and only requires 46 h compared with conventional method. Zhu et al. (2012) reported a method combining impedance technology and a rRNA lateral flow assay. In this method, 100 g of spiked or commercial PIF samples were diluted in 900 mL BPW and incubated at 30°C for 16 h, after which 0.1 mL of pre-enriched samples was transferred into measuring wells filled with 10 mg/L vancomycin. The impedance was then measured at 42°C in a BacTrac 4300 Microbiology Analyzer. The medium impedance (M-value) detects changes in the impedance caused by variations in the conductivity of the growth medium during the culture period. As soon as the threshold value (5% for the M-value) was achieved, samples were subjected to further confirmation using the RiboFlow^TMCronobacter kit. This method was found to have 85% sensitivity and 100% specificity and could complete detection within 29 h.

Many researches are dedicating to develop faster and more reliable detection method for Cronobacter species in PIF. They combined conventional detection method with other newly developed methods. For examples, Lampel and Chen (2009), Mullane et al. (2006), and Oh and Kang (2004) mingled selective media with real-time PCR, cationic-magnetic-beads and fluorescent. The incorporation of new technique with selective media made the detection time shorten from 5-7 days to 1 or 2 days. Similarly, the PCR works with other techniques such as immobilization, electrophoresis-laser induced fluorescence detector (Table 3). Immunological methods are also introduced with other methods such as liposome, strip and cross-priming amplification. The antibodies against Cronobacter species are widely cooperated with other techniques, especially biosensor, to develop more rapid and convenient detection. At the same time, new techniques, such as MALDI-TOF mass spectrometry and RNA hybridisation assay, give more ideas and inspiration for research on Cronobacter detection.

As an important opportunistic foodborne pathogen, Cronobacter species is receiving increased attention from researchers and food regulatory authorities. Cronobacter speceis are found in PIF, and there are many outbreaks and cases of the infection have been reported. USFDA and ISO method are widely used as standard method for detection of Cronobacter species although they are time consuming. Many researchers are dedicating on the development of more fast and sensitive methods for detecting Cronobacter species in PIF. Various new methods have been developed for the rapid detection of Cronobacter species, and combination of two or more of techniques have become more popular than use of a single method alone, especially combination of the conventional method with new techniques, such as selective media combined with PCR, biosensor or fluorescent. PCR can be also combined with immobilization, electrophoresis-laser induced fluorescence detector, as well as biosensors. Magnetic nanoparticles are a useful research tool of biosensor for detecting Cronobacter species, which usually composited with strip, PCR, and immunoassay. These methods enhance sensitivity, reduce detection time, and make great contributions on the PIF safety of Cronobacter species.