Enterobacter Sakazakii)

International Journal of Food Microbiology 136 (2009) 159–164

Contents lists available at ScienceDirect

International Journal of Food Microbiology

journal homepage: www.elsevier.com/locate/ijfoodmicro

Microarray-based comparative genomic indexing of the Cronobacter genus (Enterobacter sakazakii)

B. Healy a, S. Huynh b, N. Mullane a, S. O'Brien a, C. Iversen a, A. Lehner c, R. Stephan c, C.T. Parker b,⁎, S. Fanning a,⁎ a Centres for Food Safety and Food-borne Zoonomics, UCD Veterinary Sciences Centre, University College Dublin, Belﬁeld, Dublin 4, Ireland b Produce Safety and Microbiology Research Unit, Agricultural Research Service, U.S. Department of Agriculture, 800 Buchanan Street, Albany, CA 94710, USA c Institute for Food Hygiene and Safety, VetSuisse Faculty, University of Zurich, Zurich CH-, Switzerland article info abstract

Keywords: Cronobacter (Enterobacter sakazakii) is a recently defined genus consisting of 6 species. To extend our Microarray understanding of the genetic relationship between Cronobacter sakazakii BAA-894 and the other species of Comparative this genus, microarray-based comparative genomic indexing (CGI) was undertaken to determine the Genome presence/absence of genes identified in the former sequenced genome and to compare 276 selected open Cronobacter reading frames within the different Cronobacter strains. Seventy-eight Cronobacter strains (60 C. sakazakii, Enterobacter sakazakii 8 C. malonaticus,5C. dublinensis,2C. muytjensii,1C. turicensis,1C. genomospecies 1, and 1 Cronobacter sp.) representing clinical and environmental isolates from various geographical locations were investigated. Hierarchical clustering of the CGI data showed that the species grouped as clusters. The 5 C. dublinensis and 2 C. muytjensii strains examined formed distinct species clusters. Moreover, all of the C. sakazakii and 3 of 8 C. malonaticus strains formed a large cluster. The remaining C. malonaticus strains formed a sub-group within a larger cluster that also contained C. turicensis, C. genomospecies 1, and an unknown Cronobacter sp. Cronobacter sakazakii and 3 of 8 C. malonaticus strains could be distinguished from the others within the collection by the presence of 10 fimbrial related genes. Similarly, capsule and/or lipopolysaccharide (LPS) related glycosyltransferases differentiated several of the C. sakazakii strains from each other. © 2009 Elsevier B.V. All rights reserved.

1. Introduction Microarray technology facilitates the rapid analysis of an orga- nism's genome. Array-based comparative genomic indexing (CGI) has Cronobacter, a recently described genus currently consists of 6 previously been employed to examine the genetic content of a different species; C. sakazakii, C. malonaticus, C. dublinensis, collection of clinical Campylobacter jejuni strains and to assist in the C. muytjensii, C. turicensis and C. genomospecies 1(Iversen et al., identification of distinguishing features within their genomes linked 2008). This recently defined nomenclature is the result of a polyphasic to disease outcomes (Parker et al., 2007; Quinones et al., 2008). In the taxonomic investigation aimed at re-defining this group of organisms present study, microarray-based genome indexing with an array (Iversen et al., 2007, 2008). Cronobacter spp. are described as consisting of 276 ORFs was applied to 78 well-characterized Crono- opportunistic pathogens, causing bacteremia, necrotizing enterocoli- bacter isolates cultured from food, environmental and clinical sources. tis (NEC) and meningitis in immunocompromised neonates (Mullane Bacterial gene targets included most of the functional categories within et al., 2007). More recently, it has also emerged that Cronobacter spp. the genome C. sakazakii strain BAA-894 (http://genome.wustl.edu/ may cause infections among immunocompromised adults, in parti- genome.cgi?GENOME=Enterobacter%20sakazakii&SECTION=assem- cular the elderly (Gosney et al., 2006; See et al., 2007). Due to their blies). These data extend the early characterization of these opportu- ubiquitous nature, Cronobacter spp. have been isolated from a wide nistic pathogens. variety of foods (Baumgartner et al., 2009-this issue; Friedemann, 2007; Gurtler et al., 2005). While the primary reservoir of Cronobacter has yet to be defined, plant material is believed to be the likely source 2. Materials and methods (Lehner, 2009). 2.1. Bacterial strains and growth

⁎ Corresponding authors. Fanning is to be contacted at Tel.: +353 7166096; fax: All Cronobacter strains (Table 1) were cultured on Colombia blood +353 7166091. Parker, Tel.: +1 510 559 6187; fax: +1 510 559 6162. − E-mail addresses: [email protected] (C.T. Parker), [email protected] agar at 37 °C. Bacterial strains were frozen at 80 °C for long time (S. Fanning). storage.

− Table 1 ethidium bromide (10 mg ml 1). Two-hundred and seventy-six PCR Cronobacter strains used in this study. products ranging from 200 to 1000 bp were amplified from Clinical Environmental Food Unknown Water C. sakazakii strain BAA-894. All amplicons were purified on a Qiagen C. sakazakii 8000 robot using a Qiaquick 96-well Biorobot kit (Qiagen, Valencia, Canada E604, E654, CA), dried, and resuspended to an average concentration of 0.1–0.2 μg E655a, E655b, µl− 1 in 20 μl 50% dimethyl sulfoxide (DMSO) containing 0.3× saline- E656, E657, E880 sodium citrate (SSC, 1× SSC; 150 mM NaCl, 15 mM sodium citrate). All Czech E755, E756 Republic of the PCR amplicons were subsequently spotted in triplicate on Denmark E770 UltraGAPS slides (Corning, Lowell, MA) with an OmniGrid Accent France E834, E835, E836, E292, E772, (GeneMachines, Ann Arbor, MI) producing a final array that contained E837, E838, E839, E857, E859 a total of 276 features. The DNA was UV cross-linked to the microarray E840, E841, E842, slides with a Stratalinker at 300 mJ (Stratagene, La Jolla, CA). E843, E844, E845, E848, E849, E851, E852, E853 2.3. Preparation and fluorescent labeling of genomic DNA Germany E540 Holland E784, E787, E886 E305 Genomic DNA from 78 well-characterized Cronobacter strains was Ireland CFS101, CFS155, CFS109 ™ CFS174, CFS1001 isolated by using the Qiagen Dneasy kit (Qiagen, Valencia, CA) Malaysia E272, E274 according to the manufacturer's specifications. For each microarray New Zealand 254N hybridization reaction, genomic DNAs from the reference strain Russia E775 (C. sakazakii strain BAA-894) and a test strain were fluorescently UK E736 labeled with indodicarbocyanine (Cy5)-dUTP and indocarbocyanine Unknown E602 μ μ Uruguay E903 (Cy3)-dUTP, respectively. An aliquot (2 g) of DNA was mixed with 5 l USA ATCC29544, E823, E632, E821, E830 10× NEBlot labeling buffer containing random sequence octamer E824, E827, E828, E934, E935 oligonucleotides (NEB, Beverly, MA) and water to a final volume of E829, E893, E901 41 μl. This mixture was heated to 95 °C for 5 min and then cooled for

C. malonaticus 5 min on at 4 °C. After cooling, the remainder of the labeling reaction Australia E766 components were added: 5 μl of 10× dNTP labeling mix (1.2 mM each of Czech E608 dATP, dGTP,dCTP; 0.5 mM dTTP in 10 mM Tris [pH 8.0]; 1 mM EDTA), 3 μl Republin Cy3 dUTP (25 nmol) or Cy5 dUTP (25 nmol) (GE Biosciences, Piscataway, Germany E532 NJ) and 1 μl Klenow fragment of DNA polymerase I (5000 U/ml). The Ireland CFS28 CFS10 338 Unknown E685 labeling reactions were incubated overnight at 37 °C. Labeled DNA was USA E825 puriﬁed from unincorporated label using Qiaquick PCR Cleanup kits (Qiagen, Valencia, CA) and dried by vacuum. C. dublinensis Ireland E930 UK E798 2.4. Microarray hybridization USA E791 Switzerland E515 Labeled reference and test DNAs were combined in 45 μl Pronto Zimbabwe E464 cDNA hybridization solution (Corning, Corning, NY) and heated to 95 °C for 5 min. Fifteen microliters of the hybridization mixture was pipetted C. muytjensii France E603 directly onto a microarray slide and sealed with a coverslip. The USA E793 microarray slide was then placed in a hybridization chamber (Corning) and incubated at 42 °C for 18 h. Following the hybridization step, slides C. genomospecies 1 were washed twice in 2× SSC, 0.1% [w/v] sodium dodecyl sulfate at 42 °C UK E797 for 10 min, followed by two washings in 1× SSC at room temperature for C. turicensis 10 min, and ﬁnally twice in 0.2× SSC at room temperature for 5 min. The Switzerland E866 microarray slides were dried by centrifugation at 300 ×g for 10 min before scanning. At least two hybridization reactions were performed for each test strain to ensure reproducibility.

2.2. Construction of the Cronobacter DNA microarray 2.5. Microarray data analysis

DNA fragments of individual ORFs identified based on the genome Microarrays were scanned and analyzed as previously described sequence of C. sakazakii BAA-894 were amplified by using primers (Parker et al., 2006, 2007). DNA microarrays were scanned using an designed with ArrayDesigner 3.0 (Premier Biosoft, Palo Alto, CA) and Axon GenePix 4000B microarray laser scanner (Axon Instruments, Inc. purchased from MWG Biotech Inc. (High Point, NC). Each PCR reaction Union City, CA). Features and the local background intensities were (total reaction volume, 100 μl) consisted of 1× MasterAmp Taq PCR detected and quantified with GenePix 4.0 software (Axon Instruments, buffer, 1× MasterAmp Taq Enhancer, 2.5 mM MgCl2,200μM each Inc.). Poor features were excluded from further analysis if they contained dNTP, forward and reverse primers at 0.2 μM each, 0.5 U of MasterAmp abnormalities or were within regions of high fluorescent background. Taq DNA polymerase (Epicentre, Madison, WI), and approximately These data were filtered so that spots with a reference signal lower than 50 ng of genomic DNA from C. sakazakii strain BAA-894. Thermal the background plus 2 standard deviations of the background were cycling was performed with a Tetrad thermal cycler (Bio-Rad, discarded. Signal intensities were corrected by subtracting the local Hercules, CA) with the following parameters: 30 cycles of 25 s at background, and then Cy3/Cy5 ratios were calculated. To compensate for 94 °C, 25 s at 52 °C, and 2 min at 72 °C with a final extension at 72 °C unequal dye incorporation, data normalization was performed as for 5 min. Amplicons were first analyzed by conventional agarose gel described previously (Anjum et al., 2003; Parker et al., 2006). A gene (1% [w/v]) electrophoresis in 1× Tris–borate–EDTA buffer (10× TBE; was considered present when the Cy3/Cy5 (test/reference) intensity 0.89 M Tris Base, 0.89M Boric Acid, 0.02 M EDTA) stained with ratio was N0.6, as divergent when the Cy3/Cy5 intensity ratio was B. Healy et al. / International Journal of Food Microbiology 136 (2009) 159–164 161 between 0.6 and 0.3, and absent when the Cy3/Cy5 intensity ratio was Cronobacter species. The other five C. malonaticus strains formed a b0.3 (Fig. 1). These intensity ratios are similar to values used in other separate cluster distinct from the first three C. malonaticus strains microarray studies (Anjum et al., 2003; Garaizar et al., 2002; Taboada (Cluster B, Fig. 2). An additional strain that had not been speciated also et al., 2004). The presence, unknown and absence status for all genes clustered with the second group of C. malonaticus strains suggesting was converted to trinary scores (present = 2; unknown = 1; absent = that it may also be a member of this species. This strain has since been 0) as described previously by Parker et al. (2007). The CGI analysis to confirmed as a C. malonaticus strain through phenotypic profiling and assign present and absent genes to each Cronobacter strain was verified molecular subtyping. The C. dublinensis strains formed a species- with GENCOM software (Pin et al., 2006). The trinary gene scores for specific cluster and both of the C. muytjensii strains also formed a each replicate were also analyzed using the GeneSpring microarray distinct species-specific cluster. Although there was only one strain for analysis software version 7.3 (Agilent Technologies, Redwood City, CA) each of the other Cronobacter species, these species were outliers and and subjected to average linkage hierarchical clustering with the did not cluster with the other species. standard correlation similarity measure and bootstrapping of 1000 The cluster analysis also illustrated the relationship between the replicates to generate a dendrogram. The C. sakazakii CDSweregrouped CGI data and geographical origin of the certain strains. Cluster C into defined functional categories. Genes in each functional category for (Fig. 2) contains nine clinical strains from France and cluster D the C. sakazakii BAA-894 genome were analyzed against the CGI data set contains three clinical strains and two food strains also from France. with GeneSpring software. A gene that is absent from at least 2 strains in However, not all clusters were related to source and geographic origin. the collection was considered a variable gene in this context. In particular, cluster E contained strains from clinical, environmental and food sources, and from distinct geographic regions (USA, Canada, Ireland, Germany, Holland and New Zealand). 3. Results and discussion 3.2. Variable genes by functional category 3.1. Genomic comparisons of Cronobacter strains To further explore the genomic diversity among this collection of Based on the CGI analysis we assessed the CDS content of each isolates, genes were grouped by functional category. This approach Cronobacter strain relative to the sequenced C. sakazakii BAA-894 may reflect the physiological diversity among these Cronobacter strain (Supp. Table 1). In this study the DNA microarray, comprised of strains. The results showed that all Cronobacter strains in this study 276 CDS, representing approximately 10% of the BAA-894 strain possessed every gene in our data set assigned to the following genome. In total 72.5% (200 of 276) of the CDS (DNA coding sequence) functional categories: protein translation, transcription, protein and represented on the microarray were present in all Cronobacter isolates peptide secretion and energy production (Table 2). Two functional studied, defining a core set of Cronobacter genes (data not shown). categories with the most variable genes were the extracellular The genomic relationship among the Cronobacter strains was structures and the cell wall/membrane biogenesis. determined by hierarchical clustering analysis using GeneSpring (see Materials and methods). This cluster analysis demonstrated that most members of a Cronobacter species clustered together (Fig. 2). The 3.3. Species-specific genes C. sakazakii strains formed a large cluster (denoted as Cluster A in Fig. 2) that included three strains (338, CSF10 and CFS28) identified as CGI analysis identified particular genes that were identified as C. malonaticus based on the phenotypic criteria previously described species-specific, (i.e. present in all C. sakazakii but variable in other (Iversen et al., 2007, 2008). This arrangement suggests that there are Cronobacter spp.). A putative methyl-accepting chemotaxis gene distinct genetic differences between C. sakazakii and most of the other (ESA_02198) was absent in all C. dublinensis and C. muytjensii strains.

Fig. 1. Competitive genomic hybridization analysis. 162 B. Healy et al. / International Journal of Food Microbiology 136 (2009) 159–164

Fig. 2. Comparison of Cronobacter strains by cluster analysis of CGI results. An average linkage hierarchical clustering of the Cronobacter strains was compiled in GeneSpring version 7.3 from the CGI data. The branches of the dendogram are color coded for the particular Cronobacter species. Clusters A–E are described in the text. The genes are presented in the order of their positions on the genome of C. sakazakii strain BAA-894. For frames of reference, the putative ﬁmbrial gene clusters (ESA_02538–02541 and ESA_04070–04071) and the putative silver/copper resistance genes (ESA_04238–04242) are indicated F1, F2, and S, respectively. The gene status is color-coded: blue, present; light blue, variable/unknown; red, absent; gray, no data. For cutoffs of absence and presence predictions, refer to Materials and methods.

The putative fimbrial adhesion genes (ESA_02799 and ESA_03815) information is limited as the array is based on sequence information were absent in C. dublinensis, C. muytjensii, C. turicensis, and from only one sequenced strain. C. genomospecies 1 strains. Similarly, the putative fimbrial gene clusters (ESA_02538–ESA_02541 and ESA_04070–ESA_04071) were 3.4. Species-variable genes absent in C. dublinensis, C. muytjensii, C. turicensis, C. genomospecies 1 and 5 C. malonaticus strains. Capsule biosynthesis genes kpsT Certain genes and gene clusters were variable within C. sakazakii (ESA_03358) and kpsM (ESA_03359) were absent in C. dublinensis and absent from other Cronobacter spp. The putative silver/copper and C. muytjensii strains. These results suggest the possibility of using resistance gene cluster (ESA_04238–04242) was present in greater these gene targets as molecular markers for the identification of than half of the C. sakazakii strains only and was not present in any particular species within the Cronobacter genus. However, this other species. Putative fimbrial genes (ESA_01970 and ESA_01976) were present in only 14 of 60 C. sakazakii isolates. Two C. sakazakii isolates had large distinct deletions of flagellar related gene clusters (ESA_01248–01259 and ESA_01337–01352). Some genes on plasmid pESA2 were present in 3 C. sakazakii and C. turicensis strains. Genes Table 2 Variable genes within gene functional categories. identified from the array can later be evaluated as candidate markers for inclusion in a molecular-based detection protocol. Gene functional category No, variable genes/total Mullane et al. (2008) characterized the rfb locus in C. sakazakii and no. genes in category described the first serotypes, denoted as O:1 and O:2. Data from the Energy production and conversion 0/10 Cell cycle control and mitosis 0/2 CGI analysis supports the observation of Mullane et al. (2008) that Amino acid metabolism and transport 1/14 there are multiple O-antigen serotypes within the C. sakazakii species. Carbohydrate metabolism and transport 2/14 In this study, we observed that 23 of the C. sakazakii strains share a Coenzyme metabolism 2/7 common O-antigen gene cluster (Fig. 3). This observation requires Translation 0/4 further investigation as not all genes located within the O-antigen Transcription 5/31 Replication and repair 0/4 gene cluster were included on the current array. Furthermore, there is Cell wall/membrane biogenesis 13/31 the possibility that unidentified seroconversion phage may be present Cell motility 1/27 and capable of altering the serotype. Post translational modification/chaperone function 2/18 Inorganic ion transport and metabolism 7/22 Secondary metabolites 1/5 3.5. Conclusions Signal transduction 4/40 Intracellular trafficking and secretion/extracellular 20/32 In conclusion, the results from this study demonstrate the structures potential benefit of employing microarray-based CGI to explore the Defence mechanisms (efflux) 1/7 genomic content of a collection of Cronobacter strains. These results B. Healy et al. / International Journal of Food Microbiology 136 (2009) 159–164 163

Fig. 3. O-antigen cluster analysis.

highlight some of the similarities between species within this new supported by the United States Department of Agriculture, Agricul- genus and also some of the differences, based on comparison with the tural Research Service CRIS project 5325-42000-045. BAA-894 genome sequence. Genes encoding surface structures and membrane wall biogenesis genes were chosen as useful targets as they Appendix A. Supplementary data have previously been variable in E. coli and Salmonella.Ourfindings further highlight the urgent need to undertake genome sequencing of Supplementary data associated with this article can be found, in representatives of all of the Cronobacter species. Care should be taken the online version, at doi:10.1016/j.ijfoodmicro.2009.07.008. when interpreting the array results as an absence of a gene indicates no match occurred with the sequenced strain BAA-894 and does not References necessarily indicate an absence of the gene in that genome. Availability of sequence information for other species would facilitate a more Anjum, M.F., Lucchini, S., Thompson, A., Hinton, J.C., Woodward, M.J., 2003. Comparative detailed comparative genome analysis and may provide clues as to the genomic indexing reveals the phylogenomics of Escherichia coli pathogens. Infect. Immun. 71, 4674–4683. mechanisms of virulence involved in infection. This could also help Baumgartner, A., Grand, M., Liniger, M., Iversen, C., 2009. Detection and frequency of identify some of the phenotypic differences between/within species Cronobacter (Enterobacter sakazakii) in different categories of ready-to-eat foods (sugar transport and catabolism pathways for example). The array also other than infant formula. Int. J. Food Microbiol. 136, 189–192 (this issue). El-Sharoud, W., O'Brien, S., Negredo, C., Iversen, C., Fanning, S., Healy, B., 2009. presents data regarding the clustering and origins of the strain Characterization of Cronobacter recovered from dried milk and related products. collection. Although it was not the aim of this study, the data supports BMC Microbiol. 9, 24. the intention of using microarrays to epidemiologically monitor Crono- Friedemann, M., 2007. Enterobacter sakazakii in food and beverages (other than infant formula and milk powder). Int. J. Food Microbiol. 116, 1–10. bacter strains from various sources and geographical locations. Garaizar, J., Porwollik, S., Echeita, A., Rementeria, A., Herrera, S., Wong, R.M., Frye, J., Usera, Phenotypic based approaches for the identification of individual M.A., McClelland, M., 2002. DNA microarray-based typing of an atypical monophasic species may not prove to be 100% reliable. Recently more emphasis has Salmonella enterica serovar. J. Clin. Microbiol. 40, 2074–2078. been placed on molecular-based approaches. While 16S rRNA gene Gosney, M.A., Martin, M.V., Wright, A.E., Gallagher, M., 2006. Enterobacter sakazakii in the mouths of stroke patients and its association with aspiration pneumonia. Eur. J. Intern. sequencing can identify 5 Cronobacter spp., the method fails to Med. 17, 185–188. differentiate between C. sakazakii and C. malonaticus species. Sequencing Gurtler, J.B., Kornacki, J.L., Beuchat, L.R., 2005. Enterobacter sakazakii: a coliform of – of the recN gene (El-Sharoud et al., 2009; Kuhnert et al., 2009-this issue) increased concern to infant health. Int. J. Food Microbiol. 104, 1 34. Iversen, C., Lehner, A., Mullane, N., Bidlas, E., Cleenwerck, I., Marugg, J., Fanning, S., Stephan, has successfully been employed to identify all Cronobacter spp. Stoop R., Joosten, H., 2007. The taxonomy of Enterobacter sakazakii: proposal of a new genus et al. (2009-this issue) have developed a PCR method for the Cronobacter gen. nov. and descriptions of Cronobacter sakazakii comb. nov. Cronobacter differentiation of Cronobacter based on amplification of rpoB.Theresults sakazakii subsp. sakazakii, comb. nov., Cronobacter sakazakii subsp. malonaticus subsp. nov., Cronobacter turicensis sp. nov., Cronobacter muytjensii sp. nov., Cronobacter of this study suggest a whole genome array coupled with analysis of a dublinensis sp.nov.andCronobacter genomospecies 1. BMC Evol. Biol. 17, 64. diverse collection of strains encompassing multiple strains from each of Iversen, C., Mullane, N., McCardell, B., Tall, B.T., Lehner, A., Fanning, S., Stephan, R., Joosten, H., the different species could enable the development of further simplified 2008. Cronobacter gen. nov., a new genus to accommodate the biogroups of Entero- bacter sakazakii, and proposal of Cronobacter sakazakii gen. nov., comb. nov., Crono- molecular approaches to identify these bacteria at the species level and bacter malonaticus sp. nov., Cronobacter turicensis sp. nov., Cronobacter muytjensii sp. further expand our knowledge of this new genus. nov., Cronobacter dublinensis sp. nov., Cronobacter genomospecies 1, and of three subspecies, Cronobacter dublinensis subsp. dublinensis subsp. nov., Cronobacter dublinensis subsp. lausannensis subsp. nov. and Cronobacter dublinensis subsp. lactaridi subsp. nov. Int. J. Syst. Evol. Microbiol. 58, 1442–1447. Acknowledgements Kuhnert, P., Korczak, B.M., Stephan, R., Joosten, H., Iversen, C., 2009. Phylogeny and prediction of genetic similarity of Cronobacter and related taxa by multilocus The authors acknowledge the financial support provided by the sequence analysis (MLSA). Int. J. Food Microbiol. 136, 152–158 (this issue). Lehner, A. Highlighting environmental reservoir aspects for Cronobacter spp. Oral Irish Government through the Food Institution Research Measure presentation at the 1st International Meeting on Cronobacter (Enterobacter sakazakii), (FIRM) grant 05/R&D/D/363 to S.F. C.T. Parker and S. Huynh were O'Reilly Hall, University College Dublin, Ireland, 22nd–23rd January 2009. 164 B. Healy et al. / International Journal of Food Microbiology 136 (2009) 159–164

Mullane, N., Iversen, C., Healy, B., Walsh, C., Whyte, P., Wall, P.G., Quinn, T., Fanning, S., Quinones, B., Guilhabert, M.R., Miller, W.G., Mandrell, R.E., Lastovica, A.J., Parker, C.T., 2007. Enterobacter sakazakii: an emerging bacterial pathogen with implications for 2008. Comparative genomic analysis of clinical strains of Camplyobacter jejuni from infant health. Minerva Pediatr. 59, 137–148. South Africa. PLoS ONE 3, e2015. Mullane, N., O'Gaora, P., Nally, J.E., Iversen, C., Whyte, P., Wall, P.G., Fanning, S., 2008. See, K.C., Than, H.A., Tang, T., 2007. Enterobacter sakazakii bacteremia with multiple splenic Molecular analysis of the Enterobacter sakazakii O-antigen gene locus. Appl. Environ. abscesses in a 75-year-old woman: a case report. Age Ageing 37, 595–596. Microbiol. 74, 3783–3794. Stoop, B., Lehner, A., Iversen, C., Fanning, S., Stephan, R., 2009. Development and evaluation Parker, C.T., Quiñones, B., Miller, W.G., Horn, S.T., Mandrell, R.E., 2006. Comparative genomic of an rpoB based PCR system to differentiate the six proposed species within the genus analysis of Campylobacter jejuni strains reveals diversity due to genomic elements Cronobacter. Int. J. Food Microbiol. 136, 165–168. similar to those present in C. jejuni strain RM1221. J. Clin. Microbiol. 44, 4125–4135. Taboada, E.N., Acedillo, R.R, Carrillo, C.D., Findlay, W.A., Medeiros, D.T., Mykytczuk, O.L., Parker, C.T., Miller, W.G., Horn, S.T., Lastovica, A.J., 2007. Common genomic features of Roberts, M.J., Valencia, C.A., Farber, J.M., Nash, J.H., 2004. Large-scale comparative Campylobacter jejuni subsp. doylei strains distinguish them from C. jejuni subsp. genomics meta-analysis of Campylobacter jejuni isolates reveals low level of genome jejuni. BMC Microbiol. 7, 50. plasticity. J. Clin. Microbiol. 42, 4566–4576. Pin, C., Reuter, M., Pearson, B., Friis, L., Overweg, K., Baranyi, J., Wells, J., 2006. Comparison of different approaches for comparative genetic analysis using microarray hybridization. Appl. Microbiol. Biotechnol. 72, 852–859.