Comparative genomics reveals an SNP potentially leading to phenotypic diversity of Salmonella enterica serovar Enteritidis.

An SNP is a spontaneous genetic change having a potential to modify the functions of the original genes and to lead to phenotypic diversity of bacteria in nature. In this study, a phylogenetic analysis of Salmonella enterica serovar Enteritidis, a major food-borne pathogen, showed that eight strains of S. Enteritidis isolated in South Korea, including FORC_075 and FORC_078, have almost identical genome sequences. Interestingly, however, the abilities of FORC_075 to form biofilms and red, dry and rough (RDAR) colonies were significantly impaired, resulting in phenotypic differences among the eight strains. Comparative genomic analyses revealed that one of the non-synonymous SNPs unique to FORC_075 has occurred in envZ, which encodes a sensor kinase of the EnvZ/OmpR two-component system. The SNP in envZ leads to an amino acid change from Pro248 (CCG) in other strains including FORC_078 to Leu248 (CTG) in FORC_075. Allelic exchange of envZ between FORC_075 and FORC_078 identified that the SNP in envZ is responsible for the impaired biofilm- and RDAR colony-forming abilities of S. Enteritidis. Biochemical analyses demonstrated that the SNP in envZ significantly increases the phosphorylated status of OmpR in S. Enteritidis and alters the expression of the OmpR regulon. Phenotypic analyses further identified that the SNP in envZ decreases motility of S. Enteritidis but increases its adhesion and invasion to both human epithelial cells and murine macrophage cells. In addition to an enhancement of infectivity to the host cells, survival under acid stress was also elevated by the SNP in envZ. Together, these results suggest that the natural occurrence of the SNP in envZ could contribute to phenotypic diversity of S. Enteritidis, possibly improving its fitness and pathogenesis.

Data Summary

The whole genome sequences used in this study have been deposited previously in the National Center for Biotechnology Information RefSeq database (https://www.ncbi.nlm.nih.gov/), and the accession numbers are listed in Table S3. The authors confirm that all supporting data and protocols have been provided within the article and through the supplementary data files.

An SNP is an evolutionary event which contributes to phenotypic diversity of bacteria in nature. Although eight strains of serovar Enteritidis, a major food-borne pathogen, had almost identical genome sequences, differential phenotypes of biofilm formation and colony morphology were observed among the eight strains. A non-synonymous SNP in envZ, encoding a sensor kinase of the EnvZ/OmpR two-component system, was identified to be responsible for the observed differential phenotypes. The SNP in envZ affected EnvZ function, increasing the phosphorylated status of OmpR in S. Enteritidis and altering the expression of the OmpR regulon. In addition, the SNP in envZ significantly differentiated the virulence-related phenotypes of S. Enteritidis, including motility, adhesion and invasion to host cells, and even acid resistance. These results suggest that the spontaneous SNP in envZ could serve as a pathoadaptive mutation of S. Enteritidis, potentially leading to its phenotypic diversity.

Introduction

Bacterial pathogens have evolved genetically to adapt to various environmental conditions [1]. Many studies have revealed that genetic mutations occur naturally in bacteria for their optimal fitness and successful pathogenesis during the course of infection [2, 3]. Horizontal gene transfer, one of the possible mechanisms for genetic evolution, results in the acquisition of novel genes and enables non-pathogenic bacteria to be pathogenic [4]. Alternatively, small genetic changes such as SNPs have potential to modify the function of original genes and to induce phenotypic diversity of pathogens, leading to their pathoadaptive evolution [5]. Along with the rapid development of next-generation sequencing technologies, comparative genomic analysis has discovered a number of SNPs occurring in many pathogenic bacteria including , and [6-8]. However, studies on the association of each SNP with bacterial pathogenesis are still limited.

serovar Enteritidis is a major food-borne pathogen, which causes diseases ranging from mild gastroenteritis to severe systemic infection [9, 10]. has multiple two-component systems to recognize diverse environmental changes and to respond appropriately [11]. The EnvZ/OmpR two-component system has been well studied in . The sensor kinase EnvZ phosphorylates itself and transfers the phosphoryl group to its cognate response regulator OmpR in response to specific environmental signals such as osmolarity and pH change [12-14]. Phosphorylation induces conformational changes of OmpR, which enhances its DNA binding affinity [15], and thus the expression of the OmpR regulon is mostly dependent on the amount of phosphorylated OmpR (OmpR-P). The OmpR regulon includes ompF and ompC, which encode outer membrane porins, and regulation of their expression is well characterized. A low level of OmpR-P is enough to activate the transcription of ompF, but not of ompC [16]. In contrast, a high level of OmpR-P activates ompC transcription, while repressing ompF transcription [16].

The EnvZ/OmpR two-component system also regulates virulence-related genes and plays an essential role in the pathogenesis of [17, 18]. For example, the EnvZ/OmpR system controls the expression of csgD, which encodes a master regulator activating the biosynthesis of curli fimbriae and cellulose [19-21]. Curli fimbriae and cellulose are major biofilm components and contribute to the development of a red, dry and rough (RDAR) colony morphology [22-24]. In addition, the EnvZ/OmpR system represses expression of flagellar genes such as fljB and fliC, which are associated with the motility of [25]. The EnvZ/OmpR system positively regulates the expression of hilA, encoding a major activator of virulence genes located in the pathogenicity island 1 (SPI-1), and the expression of ssrA and ssrB, encoding a master activator of the SPI-2 genes [26-28]. The SPI-1 and SPI-2 encode type III secretion systems and enable the efficient invasion to host cells and intracellular replication [29-32]. Furthermore, OmpR itself is an acid shock protein and regulates the expression of genes that are necessary for acid resistance [14, 33–35].

In the present study, we conducted a phylogenetic analysis of 241 strains of S. Enteritidis and revealed that eight strains isolated in South Korea, including FORC_075 and FORC_078, have almost identical genome sequences. Interestingly, however, distinct phenotypes were observed in FORC_075 which showed an impaired biofilm formation and a smooth and white (SAW) colony morphology. We identified that nine non-synonymous SNPs have occurred exclusively in FORC_075, and one of which is in envZ resulting in an amino acid change from Pro248 (CCG) in other strains including FORC_078 to Leu248 (CTG) in FORC_075. By exchanging the SNP allele in envZ between FORC_075 and FORC_078, we demonstrate that the SNP in envZ determines the biofilm formation and colony morphology. The SNP in envZ induces functional modification of EnvZ, increasing the cellular level of OmpR-P in S. Enteritidis and altering the expression of the OmpR regulon. Further phenotypic analyses revealed that the SNP in envZ decreases motility, but increases both adhesion and invasion to host cells and elevates survival under acid stress. Together, these results suggest that the natural occurrence of the SNP in envZ plays a critical role in differentiating the virulence-related phenotypes of S. Enteritidis and thus contributes to its phenotypic diversity.

Methods

Strains, plasmids and culture conditions

The strains and plasmids used in this study are listed in Table S1 (available in the online version of this article). Unless otherwise noted, all strains were grown aerobically in Luria-Bertani (LB) medium at 37 °C. When required, antibiotics were added to the medium at the following concentrations: ampicillin, 100 µg ml−1; kanamycin, 100 µg ml−1; and chloramphenicol, 20 µg ml−1. Bacterial growth was monitored spectrophotometrically at 600 nm (A 600).

Comparative genomic analyses

The eight whole genome sequences of the S. Enteritidis strains, FORC_007, FORC_019, FORC_051, FORC_052, FORC_056, FORC_074, FORC_075 and FORC_078, were retrieved from the NCBI RefSeq database (https://www.ncbi.nlm.nih.gov/). Additionally, the 233 whole genome sequences, which were analysed at ‘complete’ and ‘chromosomal’ levels and named as ‘ subsp. enterica serovar Enteritidis’, were retrieved from the NCBI RefSeq database. All accession numbers used in this study are listed in Table S3.

For average nucleotide identity (ANI) analysis, ANI values were calculated by the JSpecies program with the blast algorithm comparing 1020 bp fragmented whole genome sequences of the eight FORC strains [36]. For phylogenetic analysis, each genome of a total of 241 strains of S. Enteritidis was mapped to the genome of S. Enteritidis P125109 (reference genome) by the Snippy program (https://github.com/tseemann/snippy). The recombinant region was removed from the resulting alignment by the Gubbins program [37], and then core SNPs were extracted by the SNP-sites program [38]. The extracted SNPs were used to calculate SNP distances between the S. Enteritidis strains by the snp-dists program (https://github.com/tseemann/snp-dists). A total of 7154 SNP sites of 241 strains were used to reconstruct a phylogenetic tree by the RAxML program with a gamma distribution and a general time-reversible model under 500 bootstrap repeats [39].

For pangenome analysis, the whole genomes of the eight FORC strains were annotated by the Prokka program [40], and their pangenome was built by the Roary program [41] using the resulting annotations. For SNP analysis, the whole genome sequences of FORC_007, FORC_019, FORC_051, FORC_052, FORC_056, FORC_074 and FORC_075 were aligned to that of FORC_078 by the NUCmer program [42]. SNP positions were inferred by show-snps programs [42], and insertions and deletions were excluded. The effect of non-synonymous SNPs on protein function was predicted by the PROVEAN (Protein variation effect analyzer, http://provean.jcvi.org/) [43], SNAP2 (Screening for non-acceptable polymorphisms 2, https://www.rostlab.org/services/snap/) [44] and SIFT (Sorting intolerant from tolerant, https://sift.bii.a-star.edu.sg/) [45] programs.

Biofilm formation

Biofilms of the S. Enteritidis strains were formed as described previously [46] with minor modifications. Briefly, overnight cultures of the S. Enteritidis strains were diluted to an A 600 of 0.01 in tryptic soy broth (1 : 20 diluted TSB), and 200 µl of the resulting culture was used to form biofilms on each well of 96-well polystyrene microtitre plates (Nunc). After static incubation at 30 °C for 24 or 48 h, the planktonic cells were removed, and the remaining biofilms were stained with 1 % crystal violet (CV) solution (Sigma-Aldrich) and quantified as described previously [47].

Colony morphology assay

For analysis of colony morphology, 1 µl of overnight cultures of the S. Enteritidis strains was used to spot onto LB agar plates without salt, containing 40 µg ml−1 of Congo red (Sigma) and 20 µg ml−1 of Coomassie brilliant blue (Sigma) (CR agar plates) or 200 µg ml−1 of calcofluor white (Sigma) (CFW agar plates). The resulting colonies were grown at 26 °C for 96 h and visualized using a Stemi 305 stereomicroscope (Zeiss) equipped with an Axiocam 105 colour camera (Zeiss) or photographed using a digital camera (PowerShot G7X Mark II; Canon).

Generation of a zirT mutant

The zirT gene (FORC78_1136) was inactivated by deletion (1465 bp of 1983 bp) of the coding region using the lambda red recombination method [48]. Briefly, a linear DNA fragment containing a kanamycin resistance (Kmr) cassette was amplified from pKD13 using ZIRT01-F and ZIRT01-R which were designed to carry 5′- and 3′-flanking regions of zirT (Table S2). The resulting fragment was introduced into the zirT coding region of FORC_078 carrying pKD46 to generate FORC_078-zirT::kan (Table S1). The deletion of zirT was confirmed by PCR.

Generation of single nucleotide substitution mutants

For single nucleotide substitutions (C → T in envZ of FORC_078 and T → C in envZ of FORC_075), the envZ genes of FORC_078 and FORC_075 were first replaced with a Kmr cassette and chloramphenicol resistance (Cmr) cassette, respectively, using the lambda red recombination method [48]. Briefly, pairs of primers, ENVZ01-F and ENVZ01-R1 or ENVZ01-F and ENVZ01-R2, which were designed to carry 5′- and 3′-flanking regions of envZ, were used for amplification of linear DNA fragments containing the Kmr cassette from pKD13 and the Cmr cassette from pKD3, respectively (Table S2). The resulting Kmr cassette was introduced into the envZ gene of FORC_078 carrying pKD46 to generate FORC_078-envZ::kan, and the resulting Cmr cassette was introduced into the envZ gene of FORC_075 carrying pKD46 to generate FORC_075-envZ::cat (Table S1).

Then, the envZ region of FORC_078 containing the SNP allele (C) and that of FORC_075 containing the SNP allele (T) were amplified using ENVZ02-F and ENVZ02-R (Table S2), and the resulting fragments were ligated into SphI-SacI-digested pCVD442 to generate pDH1903 and pDH1904, respectively (Table S1). S17-1 λpir containing pDH1903 was used as a conjugal donor to FORC_075-envZ::cat to generate FORC_075-EnvZL248P (Table S1). S17-1 λpir containing pDH1904 was used as a conjugal donor to FORC_078-envZ::kan to generate FORC_078-EnvZP248L (Table S1). The conjugation and isolation of the transconjugants were conducted using the method described previously [49]. The single nucleotide substitution was confirmed by DNA sequencing. The same experimental procedures were adopted for single nucleotide substitution (C → T) in envZ of ATCC 13076 to generate ATCC 13076-EnvZP248L (Table S1).

Purification of OmpR and Western blot analysis

The ompR gene was amplified using OMPR01-F and OMPR01-R (Table S2), and the resulting fragment was subcloned into pET-28a(+) (Novagen) to generate pDH2003 (Table S1). The His6-tagged OmpR was expressed in BL21(DE3) and purified by affinity chromatography (Qiagen). The purified His6-tagged OmpR was used to raise mouse anti-OmpR polyclonal antibody (AbClon).

For Western blot analysis, the S. Enteritidis strains grown to an A 600 of 2.5 were harvested by centrifugation, and the cells were lysed using B-PER Bacterial Protein Extraction Reagent with Enzymes (Thermo Fisher Scientific). The cell debris was removed by centrifugation to obtain clear cell lysates. OmpR and DnaK in the clear cell lysates were detected by Western blot analysis using mouse anti-S. Enteritidis OmpR antibody and mouse anti- DnaK antibody (Enzo Life Science) as described previously [50]. The phosphorylated status of OmpR was detected by Western blot analysis using 10 % SuperSep Phos-tag precast gels (Wako). After electrophoresis, the precast gels were washed three times with transfer buffer (25 mM Tris, 192 mM glycine, 20 % methanol and 5 mM EDTA) to remove Zn2+ and further washed once with transfer buffer without EDTA. The phosphorylated- and unphosphorylated-OmpR were detected using the same mouse anti-S. Enteritidis OmpR antibody.

RNA purification and transcript analysis

Total RNAs were isolated from the S. Enteritidis strains grown to an A 600 of 2.5 by using an RNeasy mini kit (Qiagen). For quantitative reverse transcription-PCR (qRT-PCR), cDNA was synthesized from 1 µg of the total RNAs by using an iScript cDNA synthesis kit (Bio-Rad). Real-time PCR amplification of the cDNA was performed by using a CFX96 real-time PCR detection system (Bio-Rad) with pairs of specific primers (Table S2) as described previously [51]. Relative expression levels of each gene were calculated by using the 16S rRNA expression level as the internal reference for normalization.

Motility, adhesion, invasion and acid resistance assay

For motility assays, 2 µl of the S. Enteritidis strains grown to an A 600 of 2.5 was used to stab into LB semisolid medium containing 0.3 % agar. The plates were incubated at 37 °C for 7 h, and the migration area of cells was visualized by a Gel Doc EZ Imager (Bio-Rad).

For adhesion assays, HeLa human epithelial cells and RAW 264.7 murine macrophage cells were grown in Dulbecco's modified Eagle medium (DMEM) (Gibco-BRL) supplemented with 10 % FBS, 100 U ml−1 penicillin and 100 µg ml−1 streptomycin. One day before bacterial infection, the HeLa cells and RAW 264.7 cells were seeded into 24-well tissue culture plates at a concentration of 2.5×105 cells per well and incubated at 37 °C under 5 % CO2. Each well was infected with the S. Enteritidis strains grown to an A 600 of 2.5 at an m.o.i. of 10, centrifuged immediately at 500 for 5 min, and then incubated for 30 min. The wells were washed three times with PBS to remove non-adherent bacteria and then lysed in 1 % Triton X-100 for 30 min. For invasion assays, the wells were further incubated for 30 min with DMEM supplemented with 100 µg ml−1 gentamicin to kill extracellular bacteria before lysis with 1 % Triton X-100. The adhered and intracellular bacteria were diluted in PBS and plated on LB agar to enumerate the c.f.u.

For acid resistance assays, the S. Enteritidis strains grown to an A 600 of 2.5 were washed once with PBS (acid-unadapted bacteria) or further incubated in M9 minimal medium containing 10 mM glucose (M9G) (pH 4.3) at 37 °C for 2 h (acid-adapted bacteria). Then, the acid-unadapted and acid-adapted bacteria were incubated in M9G (pH 3.0) at 37 °C for 2 h. The pH of M9G was adjusted with HCl. Aliquots of the resulting cultures were diluted in PBS and plated on LB agar to enumerate the c.f.u.

Sequence analysis

The protein ID of FORC_075 EnvZ (WP_080165161.1) was submitted to the NCBI Identical Protein Groups database, and a list of strains expressing the same FORC_075 EnvZ was retrieved. The isolation information of each strain was retrieved from the NCBI BioSample database, and its serovar was predicted by the SISTR (Salmonella in silico typing resource) program [52] using the whole genome sequences. All accession numbers are listed in Table S5.

Statistical analysis

Statistical analyses were performed as indicated in the figure legends using GraphPad Prism 7.0 (GraphPad Software). The significance of differences between experimental groups was accepted at a p value of <0.05.

Results

The S. Enteritidis strains isolated in South Korea have a close genetic relationship

Previously, eight strains of S. Enteritidis isolated from different sources in South Korea were collected and designated as FORC_007, FORC_019, FORC_051, FORC_052, FORC_056, FORC_074, FORC_075 and FORC_078, and their whole genomes were then completely sequenced and deposited in the NCBI GenBank database under accession numbers GCA_001305235.1, GCA_001705055.1, GCA_002313085.1, GCA_002220345.1, GCA_002313105.1, GCA_003515965.1, GCA_003429365.1 and GCA_004135835.1, respectively. To examine their genetic similarity, ANI values were first calculated. Remarkably, the ANI values between each genome were very high, ranging from 99.98 % to 100 %, which indicates that the eight FORC strains possess almost identical genome sequences. Then, to determine their evolutionary relationship in the context of the S. Enteritidis strains isolated in different countries, an SNP-based phylogenetic analysis was performed. Although all FORC strains were isolated from different sources or in different years (Table S4), they clustered very closely in the phylogenetic tree (Fig. 1), indicating again that their genetic backgrounds are highly similar. In more detail, the FORC_052, FORC_056, FORC_075 and FORC_078 strains were located in the same branch (Cluster-A in Fig. 1), and the other FORC strains, FORC_007, FORC_019, FORC_051 and FORC_074, clustered together with seven strains isolated in Asia (two from South Korea and five from China) (Cluster-B in Fig. 1). As can be inferred from the phylogenetic tree, the maximum SNP distances within Cluster-A, within Cluster-B, and between the two clusters were 49, 84 and 94, respectively (Fig. S1). The close phylogenetic relationship between the eight FORC strains and the seven Asian strains may result from their relative geographical proximity. Together, these results indicate that the eight FORC strains share their genomic features and have highly similar genetic backgrounds.

Fig. 1.

Phylogenetic tree of the S. Enteritidis strains. The phylogenetic relationships were calculated by the RAxML program [39] and visualized using the iTOL program (https://itol.embl.de/) [82]. The branch containing the S. Enteritidis strains isolated in South Korea is shown in a dashed box on the right at finer scale. Regions of isolation are indicated in different colour. Bars represent the nucleotide substitutions per site.

FORC_075 exhibits an impaired biofilm formation and a SAW colony morphology

Because the ability of to form biofilms is important for persistence and survival under environmental stresses [53-56], the biofilm-forming abilities of the eight FORC strains were evaluated. Interestingly, the biofilm-forming ability of FORC_075 was much lower than those of the other FORC strains (Fig. 2a). In particular, the amount of biofilm formed by FORC_075 was approximately 20-fold lower than that formed by FORC_078, a strain phylogenetically closest to FORC_075 (Figs 1 and 2a). One possible hypothesis for the impaired biofilm formation is that FORC_075 has lost the ability to produce the major biofilm components such as curli fimbriae and cellulose. Because both curli fimbriae and cellulose contribute to the development of the RDAR colony morphology [24], the S. Enteritidis strains were grown on agar plates containing Congo red (CR agar plates), and their colony morphologies were also compared. FORC_007 formed red and dry colonies with concentric rings (Fig. 2b), which indicates the production of curli fimbriae only [57]. While FORC_019, FORC_051, FORC_052, FORC_056, FORC_074 and FORC_078 formed RDAR colonies, FORC_075 formed SAW colonies, a significantly distinct colony morphology from those of other FORC strains (Fig. 2b). These results support our hypothesis that FORC_075 produces low levels of curli fimbriae and cellulose. To further confirm the cellulose production, the S. Enteritidis strains were grown on agar plates containing calcofluor white (CFW agar plates), and their colonies were observed under UV light [23]. Consistent with the results of Fig. 2(b), low fluorescence intensities were observed in the colonies of FORC_007 and FORC_075 compared with those of other FORC strains forming RDAR colonies (Fig. S2), indicating that FORC_007 and FORC_075 produce a small amount of cellulose. Together, these results suggest that the impaired biofilm formation of FORC_075 is due to low levels of both curli fimbriae and cellulose production. Because the genome sequences of the eight FORC strains were almost identical, the significantly different phenotypes of biofilm formation and colony morphology of FORC_075 were unexpected.

Fig. 2.

Distinct biofilm formation and colony morphology of FORC_075. (a) Biofilms of the S. Enteritidis strains were grown on 96-well microtitre plates for 48 h and quantified using CV staining. Error bars represent the standard deviation from three independent experiments. Statistical significance was determined by multiple comparisons after one-way ANOVA. ***, P<0.0005; ****, P<0.0001; ns, not significant. (b) The S. Enteritidis strains were spotted onto CR agar plates and incubated for 96 h. The colony morphology was visualized using a stereomicroscope (Stemi 305; Zeiss) at 4× magnification. Bars, 2 mm.

A single SNP in envZ is responsible for the distinct phenotypes of FORC_075

To elucidate the genetic basis for the unexpected phenotypes of FORC_075, the whole genome sequence of FORC_075 was compared with those of the other FORC strains including FORC_078 using various bioinformatics tools. First, the pangenome of the eight FORC strains was built to identify genes carried differently in the FORC strains. The size of the pangenome was 4767 genes, of which about 91 % (4364 genes) constitute the core genome with a length of 4 052 928 bp, indicating again that the eight FORC strains have similar genetic backgrounds. Among the accessory genome, a total of 10 adjacent genes (FORC78_1133 to FORC78_1142) were not detected just in FORC_075 (Fig. S3a). This region belongs to a genomic island known as GEI 1664/1678 [58], suggesting that genomic rearrangement may have occurred in the FORC_075 genome. In particular, the zir operon (FORC78_1134 to FORC78_1137; zirRTSU) is conserved throughout the serovars [59]. The zirT gene encodes a membrane transporter ZirT, which mediates secretion of ZirS and ZirU, and this secretion system plays a role as an antivirulence modulator during infection [59, 60]. Because a previous study suggested that the ZirT-dependent secretion system may play a potential role in biofilm formation [60], the zirT gene was deleted in the FORC_078 genetic background, and the biofilm- and RDAR colony-forming abilities were determined. However, the biofilm formation and colony morphology of the isogenic zirT mutant of FORC_078 were similar to those of the parent strain (Fig. S3b,c). This result indicates that loss of the zir operon in the FORC_075 genome is not responsible for its impaired biofilm formation and SAW colony morphology.

Next, the whole genome sequences of the eight FORC strains were compared at the single nucleotide level. Among a total of 198 SNPs detected in the eight strains, nine non-synonymous SNPs were unique to the FORC_075 strain (Table 1). In particular, the two genes containing SNPs, dps and envZ, are related to biofilm formation [61, 62]. The SNP in dps resulted in an amino acid change from Glu142 to Ala142 in Dps, and the SNP in envZ resulted in an amino acid change from Pro248 to Leu248 in EnvZ (Table 1). When the effect of each SNP on protein function was predicted in silico, the SNP in dps was predicted not to have significant effects on the function of Dps (Table 1). In contrast, the SNP in envZ was predicted to affect the function of EnvZ (Table 1). Accordingly, the SNP in envZ rather than the SNP in dps was considered as the most likely candidate responsible for the impaired biofilm formation and SAW colony morphology of FORC_075.

Table 1.

Non-synonymous SNPs unique to strain FORC_075

Position in FORC_078 (nucleotide)*	Position in FORC_075 (nucleotide)†	Amino acid change	Gene	Function	PROVEAN score (prediction)	SNAP2 score (prediction)	SIFT score (prediction)‡
467 447 (G)	1 806 812 (A)	E113K	phsA	Thiosulfate reductase	1.142 (neutral)	−61 (neutral)	1.00 (tolerated)
943 379 (C)	2 282 762 (T)	G186S	–	Peptidase	−5.526 (deleterious)	60 (effect)	n/a
949 118 (C)	2 288 501 (T)	P176L	bioD	ATP-dependent dethiobiotin synthetase	−9.206 (deleterious)	59 (effect)	0.00 (affect protein function)
1 754 770 (A)	3 084 775 (C)	E142A	dps	Non-specific DNA-binding protein Dps	−0.073 (neutral)	−12 (neutral)	0.27 (tolerated)
2 115 383 (C)	3 445 390 (T)	A139T	ybaO	HTH-type transcriptional regulator YbaO	−3.746 (deleterious)	59 (effect)	0.14 (tolerated)
2 331 756 (G)	3 661 765 (A)	P161S	–	Chitinase	−3.649 (deleterious)	33 (effect)	0.01 (affect protein function)
2 704 053 (A)	4 034 070 (G)	H109R	hypT	HOCl-specific transcription factor HypT	−7.449 (deleterious)	72 (effect)	0.59 (tolerated)
2 769 014 (G)	4 099 014 (A)	S510N	nrdD	Ribonucleotide reductase of class III (anaerobic), large subunit	1.705 (neutral)	−97 (neutral)	0.59 (tolerated)
3 757 731 (C)	380 056 (T)	P248L	envZ	Osmolarity sensory histidine kinase EnvZ	−9.653 (deleterious)	73 (effect)	0.00 (affect protein function)

*The nucleotide position in the chromosome of FORC_078 (RefSeq assembly accession number GCF_004135835.1).

†The nucleotide position in the chromosome of FORC_075 (RefSeq assembly accession number GCF_003429365.1).

‡N/A, not analysed by the server.

–

–

To verify the effects of the SNP in envZ on the distinct phenotypes of FORC_075, the SNP alleles of FORC_078 and FORC_075 were exchanged with each other to express EnvZP248L in FORC_078 and EnvZL248P in FORC_075, and their biofilm formation and colony morphology were evaluated. The substitution of Pro248 of FORC_078 EnvZ with Leu (P248L) reduced the biofilm-forming ability to the level even lower than that of FORC_075 (Fig. 3a). In addition, the P248L substitution abolished the RDAR colony morphology of FORC_078 and led to the SAW colony morphology similar to that of FORC_075 (Fig. 3b). Although the substitution of Leu248 of FORC_075 EnvZ with Pro (L248P) did not completely restore the ability of FORC_075 to form RDAR colonies to the level comparable to that of FORC_078, the L248P substitution dramatically increased biofilm- and RDAR colony-forming abilities of FORC_075 (Fig. 3). Thus, the combined results indicate that the SNP in envZ is a major genetic change determining the ability for biofilm formation and the type of colony morphology of S. Enteritidis.

Fig. 3.

SNP in envZ responsible for the distinct biofilm phenotypes. (a) Biofilms of the S. Enteritidis strains were grown on a 96-well microtitre plate for 24 and 48 h and quantified using CV staining. Error bars represent the standard deviation from three independent experiments. Statistical significance was determined by Student’s t-test. ***, P<0.0005; ****, P<0.0001. (b) The S. Enteritidis strains were spotted onto CR agar plates and incubated for 96 h. Colony morphology was visualized using a stereomicroscope (Stemi 305; Zeiss) at 4× magnification. Bars, 2 mm. FORC_078 and FORC_075, parent strains; FORC_078-EnvZP248L, FORC_078 expressing EnvZP248L; FORC_075-EnvZL248P, FORC_075 expressing EnvZL248P.

SNP in envZ increases OmpR-P level and alters OmpR regulon expression

To determine whether the SNP in envZ does indeed affect EnvZ function, the phosphorylated status of OmpR in the S. Enteritidis strains was examined. The P248L substitution in the FORC_078 genetic background significantly increased the OmpR-P level (Fig. 4a), while the L248P substitution in the FORC_075 genetic background decreased the OmpR-P level to one not detectable by immunoblotting (Fig. 4c). Because the amount of OmpR-P governs ompF and ompC transcriptions, the effects of the SNP in envZ on the expression levels of ompF and ompC were further investigated. As expected, the P248L substitution resulting in a greater amount of OmpR-P in FORC_078 decreased ompF expression by 20-fold and increased ompC expression by almost 4-fold (Fig. 4b). Similarly, the L248P substitution resulting in a smaller amount of OmpR-P in FORC_075 increased ompF expression and decreased ompC expression (Fig. 4d). These results demonstrate that the non-synonymous SNP in envZ modifies EnvZ function, increasing the OmpR-P level in S. Enteritidis and altering the expression levels of ompF and ompC.

Fig. 4.

Effects of the SNP in envZ on the OmpR-P level and OmpR regulon expression. Total proteins and RNAs were isolated from the S. Enteritidis strains grown to an A 600 of 2.5. (a, c) The cellular levels of phosphorylated and unphosphorylated OmpR were determined by Western blot analysis using Phos-tag SDS-PAGE gels (Wako). (b, d) Transcript levels of ompF, ompC, csgD, and fliC were determined by qRT-PCR. The transcript levels in each parent strain were set as 1. Error bars represent the standard deviation from three independent experiments. Statistical significance was determined by Student’s t-test. *, P<0.05; **, P<0.005; ***, P<0.0005. FORC_078 and FORC_075, parent strains; FORC_078-EnvZP248L, FORC_078 expressing EnvZP248L; FORC_075-EnvZL248P, FORC_075 expressing EnvZL248P.

To examine the effect of the SNP in envZ on the expression of the OmpR regulon, expression of csgD and fliC in the S. Enteritidis strains were also compared. Expression of csgD in the FORC_078-EnvZP248L and FORC_075 strains showing increased OmpR-P levels were significantly lower than those in the FORC_078 and FORC_075L248P strains showing decreased OmpR-P levels, respectively (Fig. 4). The results were consistent with previous reports that a high level of OmpR-P has a repressive effect on csgD expression [19, 63]. Meanwhile, fliC expression was also reduced in the FORC_078 strain by the P248L substitution (Fig. 4b) and elevated in the FORC_075 strain by the L248P substitution (Fig. 4d). To examine the effect of the altered expression of fliC on motility, the swimming areas of the S. Enteritidis strains on a semisolid plate surface were compared. The diameter of the swimming area of the FORC_078-EnvZP248L strain was decreased to approximately 70 % of that of the FORC_078 strain (Fig. 5). Similar to the decreasing effect of the P248L substitution on motility, the FORC_075 strain was less motile than the FORC_075-EnvZL248P strain (Fig. 5). This suggests that the SNP in envZ decreases the expression level of fliC and thus results in reduced motility of S. Enteritidis. Collectively, these results indicate that functional modification of EnvZ induced by the SNP in envZ increases the phosphorylated status of OmpR and alters the expression of the OmpR regulon, leading to phenotypic changes in biofilm formation and motility of S. Enteritidis.

Fig. 5.

Effect of SNP in envZ on motility. (a) Areas of motility of the S. Enteritidis strains grown on LB plates with 0.3 % agar for 7 h. (b) Diameter of the areas of motility (mm). Error bars represent the standard deviation from three independent experiments. Statistical significance was determined by Student’s t-test. *, P<0.05; **, P<0.005. FORC_078 and FORC_075, parent strains; FORC_078-EnvZP248L, FORC_078 expressing EnvZP248L; FORC_075-EnvZL248P, FORC_075 expressing EnvZL248P.

SNP in envZ determines the virulence-related phenotypes of S. Enteritidis

To extend our understanding of the role of the SNP in envZ in S. Enteritidis pathogenesis, the effects of the exchange of the SNP allele on the virulence-related phenotypes were examined. When HeLa human epithelial cells and RAW 264.7 murine macrophage cells were infected with the S. Enteritidis strains, the adhesion of FORC_078 to the epithelial and macrophage cells was significantly increased by the P248L substitution to the level comparable to that of FORC_075 (Fig. 6a, c). Consistent with this result, the adhesion of FORC_075 to both host cells was reduced by the L248P substitution (Fig. 6a, c). In addition, although it was not possible to assess the invasion of FORC_075 because of its gentamicin resistance (Table S1), the invasion of FORC_078 to HeLa and RAW 264.7 cells was greatly increased by the P248L substitution (Fig. 6b, d). These combined results indicate that the SNP in envZ leading to EnvZL248 increases the infectivity of S. Enteritidis to host cells.

Fig. 6.

Effect of SNP in envZ on infectivity to host cells. HeLa cells and RAW 264.7 cells were infected with the S. Enteritidis strains at an m.o.i. of 10 for 30 min. (a, c) Adhesion to HeLa cells (a) and RAW 264.7 cells (c) expressed as the ratio of the number of adhering cells to the total number of cells used for infection. Relative adhesion of FORC_078 was set as 1. (b, d) Invasion to HeLa cells (b) and RAW 264.7 cells (d) expressed as the ratio of the number of intracellular cells to the total number of cells used for infection. Relative invasion of FORC_078 was set as 1. Error bars represent the standard deviation from three independent experiments. Statistical significance was determined by Student’s t-test. *, P<0.05, **, P<0.005, ***, P<0.0005. FORC_078 and FORC_075, parent strains; FORC_078-EnvZP248L, FORC_078 expressing EnvZP248L; FORC_075-EnvZL248P, FORC_075 expressing EnvZL248P.

The effect of the SNP in envZ on survival of S. Enteritidis under acid stress was also assessed. The survival of acid-unadapted and acid-adapted FORC_078 at pH 3.0 was increased by more than 2-fold by the P248L substitution (Fig. 7). Similarly, the survival of FORC_075 at pH 3.0 was reduced by the L248P substitution to the level comparable to that of FORC_078 in both acid-unadapted and acid-adapted cells (Fig. 7). These results indicate that the SNP in envZ leading to EnvZL248 enhances the acid resistance of S. Enteritidis, regardless of the previous acid adaptation. Together, the results suggest that the SNP in envZ improves S. Enteritidis pathogenesis by elevating its infectivity to host cells and survival under acid stress during the course of infection.

Fig. 7.

Effect of the SNP in envZ on survival under acid stress. The acid-unadapted and acid-adapted S. Enteritidis strains were compared for their abilities to survive under acid stress (pH 3.0). Survival was expressed as the ratio of the number of surviving cells to the number of initially inoculated cells. Error bars represent the standard deviation from three independent experiments. Statistical significance was determined by Student’s t-test. **, P<0.005; ***, P<0.0005; ****, P<0.0001. FORC_078 and FORC_075, parent strains; FORC_078-EnvZP248L, FORC_078 expressing EnvZP248L; FORC_075-EnvZL248P, FORC_075 expressing EnvZL248P.

Effects of the SNP in envZ are not dependent on a particular S. Enteritidis genetic background

To investigate whether the effects of the SNP in envZ are specific to the FORC strains, an ATCC 13076-EnvZP248L mutant was constructed using a standard strain of S. Enteritidis, ATCC 13076. The P248L substitution in the ATCC 13076 genetic background increased OmpR-P level and altered the expression of the OmpR regulon (Figs S4a, b). Moreover, the P248L substitution decreased biofilm formation of ATCC 13076, while increasing its infectivity to host cells as well as survival under acid stress (Fig. S4c–g). All these results observed in ATCC 13076 were identical to those in FORC_078, suggesting that the effects of the SNP in envZ are not dependent on a specific genetic background.

The SNP in envZ can naturally occur in other strains

To determine whether the SNP in envZ is also found in other strains, the presence of the strain expressing EnvZL248, instead of EnvZP248, was examined in the NCBI database. Nine isolates of including FORC_075 were identified to carry the same SNP in envZ (Table S5). The isolation source of BCW_2682 and SLM287 was chicken meat, and that of CFSAN083304 was cattle intestine (Table S5), indicating that the strain expressing EnvZL248 can survive in various environments. These results imply that the spontaneous SNP in envZ is not a dead-end mutation.

Discussion

Together with the accumulation of bacterial genomic data, comparative genomic analysis has allowed us to understand dynamic genetic changes leading to phenotypic differences [64, 65]. Acquisition or loss of accessory genes and small genetic changes in core genes may have a significant impact on phenotypes, which increases the virulence and survival of bacterial pathogens under a variety of environmental stresses. In the present study, we evaluated the phenotypes of eight strains of S. Enteritidis whose genetic similarity is very high (Fig. 1). Among the eight strains, however, FORC_075 exhibited distinct phenotypes of biofilm formation and colony morphology (Fig. 2). We demonstrated that an SNP in envZ is responsible for the impaired biofilm formation and SAW colony morphology of FORC_075 (Table 1, Fig. 3).

The SNP in envZ of FORC_075 resulted in EnvZL248, different from EnvZP248 in other strains including FORC_078 (Table 1). Pro248 is positioned in an H box that is well conserved in sensor kinases including EnvZ and important for its autophosphorylation and phosphotransfer to OmpR [66]. Thus, mutation in Pro248 probably affects the phosphorylated status of OmpR and thereby its activity to regulate downstream genes. In this study, we detected higher OmpR-P levels in the S. Enteritidis strains expressing EnvZL248 instead of EnvZP248 (Figs 4a, c and S4a) and also confirmed that expression of the OmpR regulon was significantly altered depending on the amount of OmpR-P (Figs 4b, d and S4b). Interestingly, while the ompR mRNA level was not significantly influenced by the SNP in envZ, the total OmpR protein level was increased by the SNP in envZ (Fig. S5). These results suggest the potential role of Pro248 in the regulation of OmpR expression at the post-transcriptional level, which remains to be studied in the future. Nonetheless, our combined results imply that this SNP has naturally occurred at a critical site in envZ that leads to changes in the phosphorylated status of OmpR and in its regulatory activity.

FORC_075 is a clinical strain isolated from human stool (Table S3) and carries the SNP in envZ leading to EnvZL248. There are several lines of evidence that genetic changes frequently occur in a sensor kinase of the signal transduction system during host–bacteria interactions, possibly affecting the bacterial pathogenic features. For example, S. Typhimurium strains, isolated from gallbladder of mouse, had a truncated mutation in envZ and showed hyper-biofilm formation [46]. For group A (GAS), signal transduction systems such as the LiaFSR three-component system and CovSR two-component system play an important role in virulence of the pathogen [67, 68]. GAS isolates, recovered from a patient, carried an SNP in the H box of sensor kinase LiaS, and this mutation decreased virulence of a GAS but increased its colonization to host cells [69]. In addition, GAS strains, isolated after mouse infection, contained an amino acid change from Pro285 of sensor kinase CovS (corresponding to Pro248 of EnvZ) to Ser, which affects its phosphatase activity altering the expression of virulence genes [70]. Along with these previous reports, it is possible to suggest that the SNP in envZ of the FORC_075 genome also occurred spontaneously during the course of host infection.

During host infection, modulation of the virulence-related phenotypes is important for S. Enteritidis to obtain optimal fitness and successful pathogenesis. In this study, we identified that the SNP in envZ of FORC_075 decreases biofilm formation and motility but elevates infectivity to host cells and acid resistance (Figs 3 and 5–7). It has been shown that loss of biofilm components could alleviate host immune responses and lead to efficient invasion of to host cells and its hypervirulence in vivo [71-74]. Similarly, flagella are required for motility but stimulate the host immune system, and thus their overexpression results in attenuation of in the mouse model [75, 76]. Accordingly, the decreased expression of biofilm components and flagella resulting from the SNP in envZ is expected to be beneficial for survival of S. Enteritidis in host environments by reducing the chances of being detected by the immune system. Furthermore, adhesion and invasion to host cells are essential for to cause infection, and induction of acid resistance enables the bacteria to survive in acidic conditions such as the stomach [77-79]. Thus, the enhanced infectivity to host cells and acid resistance resulting from the SNP in envZ could contribute to pathogenesis. Together, these phenotypic changes introduced by the SNP in envZ may confer selective advantages to the S. Enteritidis strains expressing EnvZL248.

In recent years, single nucleotide mutations underlying the clonal expansion of have been reported, suggesting their evolutionary impact. An SNP in the promoter region of the virulence gene pgtE, causing hyperinvasion of , was proposed as a genetic signature of isolates in S. Typhimurium ST313 lineage 2 [80]. In addition, conserved SNPs in multiple loci, leading to impaired biofilm formation of , were presented as strong evidence of the parallel evolution in invasive lineages [81]. According to these recent works, we expect that the SNP in envZ serves as a pathoadaptive mutation that could potentially play a role in bacterial evolution. The emergence of several strains carrying the same SNP allele in envZ also supports our expectation (Table S5).

In summary, we revealed that eight FORC strains of S. Enteritidis had almost identical genome sequences. However, FORC_075 showed impaired biofilm- and RDAR colony-forming abilities, which was distinct from other FORC strains including FORC_078. Among non-synonymous SNPs unique to FORC_075, an SNP in envZ leading to an amino acid change from Pro248 to Leu248 was identified to result in the impaired biofilm formation and SAW colony morphology of S. Enteritidis. The effects of the SNP in envZ on phenotypic changes of S. Enteritidis are summarized in Fig. 8. The SNP in envZ induced functional modification of EnvZ, which increased the cellular level of OmpR-P in S. Enteritidis and altered the expression of the OmpR regulon. The SNP in envZ led to the decrease in motility but the increase in adhesion and invasion to host cells and even in acid resistance of S. Enteritidis. Together, these results suggest that the SNP in envZ plays a key role in differentiating the virulence-related phenotypes. Considering that the EnvZ/OmpR system is highly conserved in , it could be suggested as a good target for development of broad-spectrum antivirulence agents against many pathogens. This study provides insights into the natural occurrence of an SNP that potentially contributes to phenotypic diversity of S. Enteritidis for optimal fitness and successful pathogenesis.

Fig. 8.

Proposed model for the effects of the SNP in envZ. A non-synonymous SNP in envZ of FORC_075 results in amino acid change of EnvZ from Pro248 in FORC_078 to Leu248 in FORC_075. This amino acid change modifies EnvZ function and leads to an increase in the cellular level of OmpR-P. As a result, expression of the OmpR regulon would be altered, which results in the decreases in biofilm formation and motility and the increases in infectivity to host cells and acid resistance. Together, the phenotypic changes induced by the SNP in envZ could improve the fitness and pathogenesis of S. Enteritidis during infection, increasing the selective pressure on FORC_075.

Click here for additional data file.