Multi-locus sequence typing of Salmonella enterica subsp. enterica serovar Enteritidis strains in Japan between 1973 and 2004

Salmonella enterica subsp. enterica serovar Enteritidis (S. Enteritidis) was responsible for a worldwide pandemic during the 1980s and 1990s; however, changes in the dominant lineage before and after this event remain unknown. This study determined S. Enteritidis lineages before and after this pandemic event in Japan using multilocus sequence typing (MLST). Thirty S. Enteritidis strains were collected in Japan between 1973 and 2004, consisting of 27 human strains from individual episodes, a bovine strain, a liquid egg strain and an eggshell strain. Strains showed nine phage types and 17 pulsed-field profiles with pulsed-field gel electrophoresis. All strains had homologous type 11 sequences without any nucleotide differences in seven housekeeping genes. These MLST results suggest that S. Enteritidis with the diversities revealed by phage typing and pulsed-field profiling has a highly clonal population. Although type 11 S. Enteritidis may exhibit both pleiotropic surface structure and pulsed-field type variation, it is likely to be a stable lineage derived from an ancestor before the 1980s and/or 1990s pandemic in Japan.


Findings
Salmonella enterica subsp. enterica serovar Enteritidis (S. Enteritidis) was responsible for a worldwide pandemic during the 1980s and 1990s; however, changes in the dominant lineage before and after this event remain unknown. It is also difficult to explain why multiple clones of S. Enteritidis (defined by phage types, (PTs)) infecting chicken reproductive tissues emerged simultaneously in geographically separate countries during this pandemic. For example, PT 4 emerged in some European countries during the late 1980s, including England and Wales [1], France [2] and Germany [3], while PT 6 dominated Denmark [4]. PT 8, PT 13a and PT 13 were most common in the USA [5] and in Japan, PT 1 and PT 4 were dominant [6]. To understand the simultaneous emergence of this pathogen, it is necessary to deduce whether these multiple lineages co-instantaneously acquired the ability to colonize chicken reproductive tissues or a single clone that possessed this ability evolved into multiple lineages before the pandemic.
Multi-locus sequence typing (MLST) can be used to accurately identify bacterial lineages [7,8]. The genetic distance between two strains can be quantitatively estimated as allelic differences in the nucleotide sequences of housekeeping or virulent genes among bacterial strains [8]. MLST data have been used in evolutionary and population analyses that estimate recombination and mutation rates and also in the investigation of evolutionary relationships among bacteria classified within the same genus [8]. Using MLST, our study aimed to determine the lineages of S. Enteritidis before and after the pandemic during the 1980s and/or 1990s in Fukuoka Prefecture, Japan.
Salmonella strains used in this study are listed in Table 1. Human strains from individual food-borne disease outbreaks in 1973-1981 were isolated in the Kanagawa Prefectural Institute of Public Health. Other human strains (isolated from individual outbreaks), the liquid egg strain and the eggshell strain were isolated from different samples at the Fukuoka Institute of Health and Environmental Sciences. The cattle strain was provided by the Livestock Hygiene Service Center. All strains were isolated in Japan between 1973 and 2004 (Table 1). We selected these strains to assess longterm alternation or rotation of S. Enteritidis genotypes. The study has been approved by the Fukuoka Institute of Health and Environmental Silences Ethics Committee.
Housekeeping genes, thrA (aspartokinase+homoserine dehydrogenase), purE (phosphoribosylaminoimidazole carboxylase), sucA (alpha ketoglutarate dehydrogenase), hisD (histidinol dehydrogenase), aroC (chorismate synthase), hemD (uroporphyrinogen III cosynthase) and dnaN (DNA polymerase III beta subunit) were used for MLST characterization. Each overnight-culture of the strains grown in LB broth (Difco Laboratories, Sparks, MD., USA) was centrifuged at 12,000 rpm for 5 min before the supernatant was discarded. The pellet was resuspended in normal saline (1 ml) and centrifuged at 12,000 rpm for 5 min. The supernatant was discarded and the pellet was resuspended in 200 μl of 0.1 × TE buffer before being boiled for 10 min and centrifuged at 12,000 rpm for 10 min. The supernatant was used as a template for the primary PCR. Primers for the primary PCR were as follows: thrA   Sequences were submitted to the MLST database website (http://mlst.ucc.ie/) and assigned existing or novel allele type numbers and sequence type numbers defined by the database. Composite sequence types (STs) were assigned based on the set of allele types derived from each of the seven loci. Allele sequences for each strain were then concatenated in the order aroC -dnaN -hemD -hisD -purE -sucA -thrA for a final composite length of 3,336 bp. Concatenated sequences were aligned using ClustalW software [9] and a phylogenetic tree was constructed using the neighbor-joining method (Center for Information Biology and DNA Data Bank of Japan) to compare to other published STs. Phylogenetic analyses were performed using NJplot (http://pbil.univ-lyon1.fr/software/njplot.html).
Phage typing was conducted on S. Enteritidis strains recovered from samples in this study according to the method of Ward et al. [10] at the National Institute of Infectious Disease.
Pulsed-field gel electrophoresis (PFGE) was conducted as described in our previous report [11]. Strains were considered different if variation in one or more DNA bands was observed. Using FPQuest Software (Bio-Rad Laboratories Inc. Hercules, CA, USA), similarity and cluster analyses were performed with Dice coefficients of similarity and an unweighted pair group method using average linkage, respectively.
Phage typing identified nine PTs (Table 1). PT 1 was most dominant, while PT 3, PT 7 and PT 36 were individually detected in only one strain for each. For PFGE analysis, 30 strains generated 17 pulsed-field profiles (PFPs) after restriction with BlnI (Table 1 Figure 1). PFP 1 was most dominant (11 strains), followed by PFP 10 (three strains). All other PFPs were only found in a single strain. S. Enteritidis with PFP 1 consisted of PT 1, PT 4, PT 14b and PT 36 strains.
For the first reported time, our study characterized all tested S. Enteritidis strains between 1973 and 2004 as a single ST (ST 11) and several PTs and PFPs. Two main interpretations of these results include: (A) Rabsch and colleagues speculated that S. Enteritidis acquired a niche in chicken reproductive tissues before the 1980s [12]; (B) disease-associated S. Enteritidis had a higher clonal population in Fukuoka Prefecture, Japan. These hypotheses are further discussed below.
Persistence of S. Enteritidis clonality in Japan supports the idea that the ST 11 lineage had possibly acquired a niche within chicken reproductive organs before the pandemics (1980s and 1990s). It is reported that the process of chicken oviduct colonization of S. Enteritidis is complex and depends on many factors including fimbriae, flagellae, lipopolysaccharide, cell wall structure and stress tolerance [13]. Until now, it has not been clear as to why S. Enteritidis is specifically implicated in egg contamination [14]. Housekeeping genes potentially play a less important roll in chicken oviduct colonization than genes of the abovementioned factors. However, the absence of changes in the housekeeping genes suggests that the ST11 lineage survived without any substantial genetic changes being produced by the fundamental processes of mutation, recombination, selection, drift, and migration [15]. Insufficient information was available for strains isolated before the pandemic because only three isolates were tested. Although ST 11 S. Enteritidis may produce some strain variation in pleiotropic surface structures [16] defined by PTs (such as PT 1,3,4,6a,7,9,14b, and 36) and genotypes defined by PFPs, it is likely to be a stable lineage that was derived from an ancestor before the pandemic in Japan. Our study therefore potentially supports the speculative assumptions made by Rabsch et al. [12].
Our study also suggested that food-borne diseaseassociated S. Enteritidis has a relatively higher clonal population than other Salmonella. In another study, 81 isolates of S. Newport from humans, food animals, and retail foods showed 12 STs [17], while in our present study, S. Enteritidis populations showed a single ST lineage. As all typing systems are influenced by the tested population [18], we can only discuss human and food related strains in Fukuoka Prefecture, Japan, rather than worldwide S. Enteritidis population genetics. However, according to the international MLST database, among the several distinct STs of S. Enteritidis in the world, only ST 11 consists of human strains. These results imply that human and food related-S. Enteritidis may be a higher clonal population than S. Newport.
MLST is a useful method for analyzing the core genes of pathogens that are of public health importance. MLST can be used to analyze conserved core genes, like housekeeping genes, which are universally conserved and provide fundamental genetic information [19]. Our study has revealed that MLST is suitable for evolutionary, basic relatedness or population studies, however, in contrast with PFGE, this method is not effective for epidemiological purposes. This is because MLST cannot analyze a diverse range of Salmonella strain specific genetic information that may be associated with niche adaptation. However, genetic variation in pathogen populations is a major barrier to disease control [20], and therefore, it is pertinent that we continue to survey the S. Enteritidis lineage in Japan using MLST.  Figure 1 Pulsed-field gel electrophoresis analysis of fingerprints from genetically defined Salmonella enterica subsp. enterica serovar Enteritidis strains. Numbers on the right of the dendrogram indicate pulsed-field profiles (PFPs) and the phage types (PTs) associated with each PFP. The scale indicates the percentage similarity, as determined using Dice coefficients. PFPs 1 and 10 include some strains. The center diagrams are the PFPs of Salmonella enterica subsp. enterica serovar Enteritidis strains with BlnI digestion.