Abstract Escherichia coli colonizes various body parts of animal hosts as a commensal and a pathogen. It can also persist in the external environment in the absence of fecal pollution. It remains unclear how this species has evolved to adapt to such contrasting habitats. Lysogeny plays pivotal roles in the diversification of the phenotypic and ecologic characters of E. coli as a symbiont. We hypothesized that lysogeny could also confer fitness to survival in the external environment. To test this hypothesis, we used the induced phages of an E. coli strain originating from marine sediment to infect a fecal E. coli strain to obtain an isogenic lysogen of the latter. The three strains were tested for survivorship in microcosms of seawater, marine sediment and sediment interstitial water as well as swimming motility, glycogen accumulation, biofilm formation, substrate utilization and stress resistance. The results indicate that lysogenic infection led to tractable changes in many of the ecophysiological attributes tested. Particularly, the lysogen had better survivorship in the microcosms and had a substrate utilization profile resembling the sediment strain more than the wild type fecal strain. Our findings provide new insights into the understanding of how E. coli survives in the natural environment. lysogeny, Escherichia coli, fitness INTRODUCTION Escherichia coli is a highly versatile species capable of taking up a multitude of niches in a wide range of highly contrasting habitats. In animal bodies, the primary niche of E. coli is a commensal that colonizes the mucus lining of the lower intestines while there are also strains that take on the niche as intra- or extra-intestinal pathogens (Goosney, Gruenheid and Finlay 2000; Johnson and Russo 2002; Croxen and Finlay 2010). Escherichia coli is also prevalent in the external environment as a result of host defecation and excretion (Brennan et al.2010). It is conventionally believed that E. coli as a symbiont should survive and reproduce poorly after leaving the animal hosts (Winfield and Groisman 2003). However, mounting evidence indicates that some populations of E. coli live autochthonously in the external environment without known association with fecal contamination. For example, E. coli has been found persisting in forest and riparian soils that were protected from recent fecal input (Byappanahalli et al.2006; Ishii and Sadowsky 2008). In addition, culturable E. coli was observed to leach continuously from the soils of outdoor lysimeters for over 9 years after the last adding of fecal material (Brennan et al.2010). The results of lab studies suggest that environmental matrices such as sediment, beach sand, soil and periphytons could provide conditions (e.g. moisture, surface for attachment, nutrients for growth and protection from solar radiation) that are suitable for E. coli to survive and reproduce (Badgley et al.2010; Garzio-Hadzick et al.2010; Ishii et al.2010; Moreira et al.2012). However, E. coli would still need to cope with a wide range of stresses that are qualitatively and quantitatively different from those in the animal bodies, and compete with a dense community of indigenous bacteria so as to establish a niche in the external environment. How E. coli is able to do so and how such traits have been acquired remain open questions (Savageau 1983; Whittam 1989; Ishii et al.2006). The answers to these questions are essential to the comprehensive understanding of the evolution, biogeography and ecophysiology of E. coli beyond the well-investigated lab model strains, commensal isolates and pathogenic serotypes. Lysogeny, a phage-mediated horizontal gene transfer mechanism, has contributed greatly to the genome plasticity of E. coli that underlies a vast intraspecific diversification of phenotypic and ecologic characters (Bergthorsson and Ochman 1998). The complete genome sequences of E. coli currently present in the Genbank vary by as much as 1.9 Mb in size, ranging from 4.0 to 5.9 Mb. In the pangenome identified for commensal and pathogenic E. coli strains, only 11% of the genes are associated with the core genome (Kaas et al.2012). Remarkably, prophages constituted 26% of the strain-specific genes in the accessory genome (Touchon et al. 2009). Through the addition of new genes and the regulation of bacterial genes, prophages play phenomenal roles in shaping the metabolism, physiology, pathogenicity, and hence ecological niche of E. coli in animal bodies (Kwon, Seong and Kim 2013; Wiles et al.2013). For instance, many genes encoding for factors that enable E. coli to attach and colonize animal tissues are prophage-like elements. Some of the factors such as the locus of enterocyte effacement-encoded proteins also confer pathogenicity (Le Gall et al.2007). The prophage-encoded Shiga-toxin genes mediate the evolution and diversification of the enterohemorrhagic O157:H7 (Ohnishi, Kurokawa and Hayashi 2001; Hayashi, Makino and Ohnishi 2001; Wick et al.2005). In the extraintestinal pathogenic E. coli CFT073, the prophage-borne neaT gene confers niche-specific advantages for systemic infection in a zebrafish model (Wiles et al.2013). Notably, neaT exhibited signatures of recent transfer from an extra-phyletic source, likely Bacteroidetes or Firmicutes, into the proteobacterial pangenome. Studies using model strains such as K-12 and its derivatives have also indicated prophages as important mediators of context-specific functions such as biofilm formation, energy conservation, and protection against antibiotics, osmotic and acid stress (Chen et al.2005; Wang et al.2010; Veses-Garcia et al.2015). It has been estimated that on the global scale half of the total population of naturally occurring E. coli exists outside animal bodies, although recent findings suggest that some of the strains observed are evolutionarily too distinct to be classified as E. coli (Walk et al.2009). However, the evolution, physiological character and ecological niche of the E. coli populations residing outside animal bodies remain largely unknown (Luo et al.2011). Building upon the knowledge about the important roles that prophage-encoded genes play in shaping the niche of commensal and pathogenic E. coli, we hypothesize that lysogeny is a possible mechanism to mediate the fitness of E. coli in the external environment. In this study, we investigated for the first time possible effects of lysogeny on the survivorship of E. coli in external environmental matrices. This was achieved using three E. coli strains as the model organisms, and the microcosms of autoclaved seawater and marine sediment, and filter-sterilized sediment interstitial water as the test environments. One E. coli strain was isolated from intertidal sediment and another one from pig feces. The third strain was an isogenic lysogen of the fecal strain that was infected by a phage induced from the first strain. The sediment strain survived in seawater and sediment microcosms better than the fecal strain did. More importantly, the isogenic lysogen of the fecal strain, which carried a prophage of the sediment strain, also exhibited better survivorship than the wild type fecal strain. The phenotypes of the three strains, including swimming motility, glycogen accumulation, biofilm formation, substrate utilization and stress resistance, were compared in order to determine what traits have been acquired through lysogeny. MATERIALS AND METHODS Escherichia coli strains Three E. coli strains, namely E1140, E455 and E455-L were involved in this study. Strain E1140, which belongs to phylogenetic group B1, was isolated from the surface sediment (<10 cm deep) of an intertidal mudflat at a Site of Special Scientific Interest in eastern Hong Kong with no obvious anthropogenic pollution (22.412°N, 114.275°E). Strain E455, which is affiliated with phylogenetic group A, was isolated from the freshly voided feces of pigs living in a farm in western Hong Kong (22.486°N, 114.006°E). Strain E455-L was created by infecting E455 with the phages induced from E1140 as detailed in Supplementary Information 1 (available online). The identity of E455-L as being an isogenic lysogen of E455 was confirmed by further infection assays, PCR-genomic fingerprinting and genome sequence analysis (Supplementary Information 2, available online). Whole genome sequencing analyses The draft genome sequences of E1140, E455 and E455-L, as published elsewhere (Lai et al.2014), were obtained using 300-bp insert pair-end libraries on the Illumina HiSeq2000 system (San Diego, CA, USA) and 10-kb insert single molecule, real-time (SMRT) technology on the PacBio RS system (Menlo Park, CA, USA). The reads were de novo assembled manually into contigs using Celera Assembler (Denisov et al.2008), version 8.1 (for E1140) and 8.2 (for E455 and E455-L). The assembled contigs of the three strains were concatenated separately into three single sequences for assessing genome rearrangements by Genome Pair Rapid Dotter (Krumsiek, Arnold and Rattei 2007) and Mauve software (Darling et al.2004). The concatenated sequences of E455 and E455-L were compared using Mauve to detect single nucleotide polymorphisms (SNPs) and genome modifications (deletions or insertions). The protein coding sequences (CDS) were predicted and annotated by using RAST (Aziz et al.2008). Homologous comparison of all the genes identified was performed by using BLAST with Kyoto Encyclopedia of Genes and Genomes (Kanehisa and Goto 2000) for function classification. Regions of prophage genes were identified using PHAST (Zhou et al.2011) and Prophinder (Lima-Mendez et al.2008). While PHAST requires either raw or annotated sequences, Prophinder works with annotated ones only. Tryptic soy broth growth assay The growth of E455, E455-L and E1140 in 1X tryptic soy broth (30 g L−1, TSB; Oxoid; Basingtok, UK) was determined by measuring absorbance at 600 nm (OD600). One milliliter of overnight culture was inoculated into 10 ml fresh medium and incubated at 37°C with shaking (150 rpm) until the culture reached exponential phase (OD600 = ∼0.2). After that the culture was divided into 500 μl aliquots. Each aliquot was inoculated into 5 ml fresh medium and incubated at 25, 30 and 37°C, respectively, with shaking at 150 rpm. One tube was taken for absorbance measurement every 20, 30 or 45 min after inoculation. Triplicate assays were performed. The maximum specific growth rate μmax was calculated using the steepest slope between four consecutive time points (Berney et al.2006). Microcosms The survivorship of E1140, E455 and E455-L in external environmental matrices was tested using microcosms of seawater, intertidal sediment and sediment interstitial water, respectively, all at 30°C. All microcosm experiments were conducted twice. Seawater for the microcosms was collected from the shore near the Coastal Marine Laboratory of Hong Kong University of Science and Technology. The seawater was autoclaved and filtered through 0.22 μm and 100 kDa tangential flow filtration (TFF) cartridges to remove bacteria and viruses, respectively. For each E. coli strain, triplicate overnight cultures were harvested and suspended in the filtered seawater to reach the final cell density of ∼3 × 104 colony-forming units (CFU) ml−1. Then the suspensions were incubated in a dark shaker incubator (150 rpm) at 30°C for 4 days. Cell density was checked every 12 h in the first day and every 24 h in the subsequent days. At each time point, serially diluted bacterial suspensions were filtered onto 0.45 μm membranes and incubated on tryptic soy agar (TSA; Oxoid; Basington, UK) at 37°C overnight for subsequent enumeration of colony-forming units. Sediment for microcosms was collected from the intertidal mudflat where E. coli E1140 was isolated. Escherichia coli suspensions at ∼1.5 × 104 CFU ml−1 were prepared as described above. For each E. coli strain, 1 ml cell suspension was thoroughly mixed with 50 g autoclaved sediment. The mixture was transferred into a 50 ml centrifuge tube, overlaid with 0.22 μm filtered seawater, and incubated at 30°C in the dark. Cell density in the sediment was checked every 24 h during the first 5 days; the intervals between time points continuously increased from 48 h to 168 h (7 days) over the next 68 days. At each time point, 50 g sediment was shaken in 100 ml 1% sterile peptone water for 30 min (250 rpm). Then, the supernatant was centrifuged (1000 rpm, 2 min) to remove the large particles (Craig et al.2004). Escherichia coli concentration in the supernatant was enumerated using the method described for the seawater microcosm experiment. Interstitial water was obtained by the centrifugation of intertidal sediment at 2000 rpm for 10 min. The supernatant was filtered through 0.22 μm and 100 kDa TFF cartridges to remove bacteria and viruses, respectively. The experimental procedures were the same as the seawater microcosms described above, except that seawater was replaced by interstitial water. Phenotypic tests Phenotypic tests including glycogen accumulation, swimming motility, biofilm formation and phenotypic microarray were performed using overnight cultures of the three E. coli strains in triplicate. An incubation temperature of 30°C was used for all tests. Glycogen accumulation was tested as previously described (Govons et al.1968). Escherichia coli was spotted onto M9 agar with 0.4% glucose. After 2 days of incubation, the colonies formed were stained in an iodine solution for 3 min, and the color of colonies was captured by using a camera within 1 min after the staining. The color intensity of the colonies, which is proportional to the amount of glycogen accumulated, was quantified using ImageJ (Abràmoff, Magalhães and Ram 2004). Escherichia coli E4702 and E. coli E792 were used as positive control and negative control, respectively. Swimming motility was investigated by inoculating E. coli into semisolid motility agar (Luria-Bertani broth solidified with 0.3% agar) as in a previous study (Barker et al.2004). After 1 day of incubation, the diameter of the colonies formed was measured using digital imaging and the ImageJ software. Escherichia coli K-12 MG1655 and E. coli E1728 were used as positive control and negative control, respectively. To test for biofilm formation, the culture of each E. coli strain was diluted 10-fold in fresh 0.2X TSB as well as in fresh 1X TSB. The dilution was transferred to 96-well microtiter plates and incubated for 3 days. After that, the plates were rinsed with autoclaved water and stained with 0.1% crystal violet (Sigma-Aldrich; St. Louis, MO, USA) for 10 min. Excessive stain was removed by rinsing with autoclaved water. Stain remaining in the well was dissolved in 30% acetic acid and measured for OD600 (Burton et al.2007). TSB broth without inoculation was used as a negative control while E. coli G23 was used as a positive control. The metabolic profiles of E455, E455-L and E1140 were determined using the OmniLog Phenotype Microarrays system (Biolog; Hayward, CA, USA). The E. coli strains were tested on PM1 and PM2 (carbon sources), PM9 (osmolytes) and PM10 (pH) in triplicate using the manufacturer's procedure. Bacterial growth data, which integrated the lag time, slope and maximum value of the growth curve in each test, were collected over 7 days of incubation at 30°C and analyzed using the opm R package (Vaas et al.2013). The data were normalized by an R pipeline to minimize plate-to-plate variation between replicates (Vehkala et al.2015). Statistical analyses To measure the growth of E. coli in the microcosms, the maximum specific growth rate (μmax) was calculated as the steepest slope between two time points on the population dynamics curve. For population decay, the first-order inactivation rate constant, kmax, was determined using the best-fit model in the GInaFiT package (Geeraerd, Valdramidis and Van Impe 2005). The values of μmax and kmax obtained from the three E. coli strains were first checked for normality and heterogeneity using a variance test. Data sets that passed the two tests were subsequently compared for differences using one-way ANOVA (95% confidence interval) followed by post hoc Tukey's test. RESULTS AND DISCUSSION Genome relatedness of the three E. coli strains The genome sequences of the three E. coli strains in this study and the genome sequences of E. coli K-12 shared 98.55–99.05% average nucleotide identity (ANI) for the genes that they had in common. The fecal strain E455 and its isogenic lysogen E455-L had 100% ANI whereas the ANIs between these two strains and E1140 were 98.56 and 98.57%, respectively. The comparison of the genome sequences of E455 and E455-L using a dot-plot revealed no observable chromosomal rearrangement between the two (Supplementary Information 3, available online). The visualization of whole genome alignment in the Mauve diagram also gave the same conclusion about chromosome arrangement between E455 and E455-L, while substantial differences were observed between the two strains and E1140 (Supplementary Information 4, available online). Further genome sequence analysis indicated that E445-L had 72 538 bases extra and was missing 59 890 bases in comparison to E455. These differences were mainly associated with the insertion of a P2-prophage derived from E1140 (discussed below) and the rest being contig gaps and insertion/deletion sites. The two genome sequences also had 81 nucleotide positions being identified as single nucleotide polymorphisms. At present, we are unable to determine whether the insertions/deletions and substitutions were actual genome modification or sequencing errors, unless they are further investigated by using PCR and completing their whole genome sequences. Nonetheless, the results of genome sequence analysis (further discussed below) together with those of PCR-genomic fingerprinting and infection assays (Supplementary Information 1 and 2) are in support of E455-L being an isogenic lysogen of E455. Prophages associated with the three E. coli strains E1140 was observed carrying six or seven prophage regions, depending on the method of sequence analysis being used (Table 1). Among the six regions commonly detected by all three methods were two P2-like prophages that shared 69% similarity in their DNA sequences. These six prophage regions were classified as intact by using the scoring system of PHAST (Table 1). Indeed, the DNA sequences of these regions were also found among the randomly amplified DNA of the phage particles induced from E1140, verifying their association with E1140 as inducible prophages (Supplementary Information 5, available online). In contrast, a prophage region that was detectable only by using PHAST analysis of the raw genome sequence was not observed among the randomly amplified DNA sequences of the induced phages. Table 1. Prophages associated with the three E. coli strains. Closest match (phage) Method Completeness Length (bp) CDS GC-content (%) Coverage (%) Identity (%) E1140 Enterobacteria phage phiP27 (Myoviridae) PT; raw Intact 84602 85 50.88 6 86 PT; ann. Intact 60545 76 50.73 8 86 PR; ann. N.A. 46232 59 51.51 12 86 Enterobacteria phage lambda (Siphoviridae) PT; raw Intact 38373 43 51.33 55 94 PT; ann. Intact 39343 44 51.28 53 96 PR; ann. N.A. 32267 36 52.37 66 96 Salmonella phage SPN9CC (Podoviridae) PT; raw Intact 41975 46 48.74 1 92 PT; ann. Not detected PR; ann. Not detected Enterobacteria phage mEp460 (Siphoviridae) PT; raw Intact 55817 50 50.16 50 97 PT; ann. Intact 55817 53 50.16 50 97 PR; ann. N.A. 44595 47 49.82 63 97 Enterobacteria phage P2 (Myoviridae) PT; raw Intact 39031 49 52.24 62 96 PT; ann. Intact 39031 48 52.24 62 96 PR; ann. N.A. 30504 39 52.39 78 96 Enterobacteria phage P88 (Myoviridae) PT; raw Intact 48962 47 52.32 57 96 PT; ann. Intact 48962 48 52.24 57 96 PR; ann. N.A. 21367 25 54.32 91 97 Enterobacteria phage P2a (Myoviridae) PT; raw Intact 34792 44 51.62 69 98 PT; ann. Intact 34792 47 51.62 69 98 PR; ann. N.A. 29341 40 52.24 76 98 E455 Salmonella phage SPN9CC (Podoviridae) PT; raw Incomplete 14553 19 47.59 6 94 PT; ann. Incomplete 12680 13 46.64 7 94 PR; ann. N.A. 5194 8 44.09 18 94 Enterobacteria phage HK629 (Siphoviridae) PT; raw Intact 60468 58 49.67 32 96 PT; ann. Intact 60467 58 49.23 34 96 PR; ann. N.A. 32065 37 52.32 64 96 Enterobacteria phage lambda (Siphoviridae) PT; raw Not detected PT; ann. Not detected PR; ann. N.A. 18596 16 49.44 3 95 E455-L Salmonella phage SPN9CC (Podoviridae) PT; raw Incomplete 14553 19 47.59 6 94 PT; ann. Incomplete 12680 13 46.64 7 94 PR; ann. N.A. 5194 8 44.09 18 94 Enterobacteria phage HK629 (Siphoviridae) PT; raw Intact 64848 67 49.67 32 96 PT; ann. Intact 64847 58 49.67 32 96 PR; ann. N.A. 32065 37 52.32 64 96 Enterobacteria phage P2a (Myoviridae) PT; raw Intact 34794 46 51.61 69 98 PT; ann. Intact 34793 44 51.61 69 98 PR; ann. N.A. 29341 40 52.24 76 98 Enterobacteria phage lambda (Siphoviridae) PT; raw Not detected PT; ann. Not detected PR; ann. N.A. 18596 16 49.44 3 95 Closest match (phage) Method Completeness Length (bp) CDS GC-content (%) Coverage (%) Identity (%) E1140 Enterobacteria phage phiP27 (Myoviridae) PT; raw Intact 84602 85 50.88 6 86 PT; ann. Intact 60545 76 50.73 8 86 PR; ann. N.A. 46232 59 51.51 12 86 Enterobacteria phage lambda (Siphoviridae) PT; raw Intact 38373 43 51.33 55 94 PT; ann. Intact 39343 44 51.28 53 96 PR; ann. N.A. 32267 36 52.37 66 96 Salmonella phage SPN9CC (Podoviridae) PT; raw Intact 41975 46 48.74 1 92 PT; ann. Not detected PR; ann. Not detected Enterobacteria phage mEp460 (Siphoviridae) PT; raw Intact 55817 50 50.16 50 97 PT; ann. Intact 55817 53 50.16 50 97 PR; ann. N.A. 44595 47 49.82 63 97 Enterobacteria phage P2 (Myoviridae) PT; raw Intact 39031 49 52.24 62 96 PT; ann. Intact 39031 48 52.24 62 96 PR; ann. N.A. 30504 39 52.39 78 96 Enterobacteria phage P88 (Myoviridae) PT; raw Intact 48962 47 52.32 57 96 PT; ann. Intact 48962 48 52.24 57 96 PR; ann. N.A. 21367 25 54.32 91 97 Enterobacteria phage P2a (Myoviridae) PT; raw Intact 34792 44 51.62 69 98 PT; ann. Intact 34792 47 51.62 69 98 PR; ann. N.A. 29341 40 52.24 76 98 E455 Salmonella phage SPN9CC (Podoviridae) PT; raw Incomplete 14553 19 47.59 6 94 PT; ann. Incomplete 12680 13 46.64 7 94 PR; ann. N.A. 5194 8 44.09 18 94 Enterobacteria phage HK629 (Siphoviridae) PT; raw Intact 60468 58 49.67 32 96 PT; ann. Intact 60467 58 49.23 34 96 PR; ann. N.A. 32065 37 52.32 64 96 Enterobacteria phage lambda (Siphoviridae) PT; raw Not detected PT; ann. Not detected PR; ann. N.A. 18596 16 49.44 3 95 E455-L Salmonella phage SPN9CC (Podoviridae) PT; raw Incomplete 14553 19 47.59 6 94 PT; ann. Incomplete 12680 13 46.64 7 94 PR; ann. N.A. 5194 8 44.09 18 94 Enterobacteria phage HK629 (Siphoviridae) PT; raw Intact 64848 67 49.67 32 96 PT; ann. Intact 64847 58 49.67 32 96 PR; ann. N.A. 32065 37 52.32 64 96 Enterobacteria phage P2a (Myoviridae) PT; raw Intact 34794 46 51.61 69 98 PT; ann. Intact 34793 44 51.61 69 98 PR; ann. N.A. 29341 40 52.24 76 98 Enterobacteria phage lambda (Siphoviridae) PT; raw Not detected PT; ann. Not detected PR; ann. N.A. 18596 16 49.44 3 95 GC, guanine-cytosine; N.A., not applicable. Raw sequences (raw) or annotated (ann.) sequences of the three genomes were analyzed by two web-based prophage region-predicting tools: PHAST (PT) and Prophinder (PR). The predicted prophage regions were BLAST searched against a phage database and results of the closest match are shown. Completeness of the prophage was assigned by PHAST according to a scoring system. CDS corresponds to the number of coding sequences predicted by PHAST or Prophinder. a indicates the homologous prophage sequence in both E1140 and E455-L. View Large When the draft genome sequence of E455 (raw or annotated) was analyzed using PHAST, it was observed carrying two prophage regions (Table 1). One was an intact region closely related to enterobacteria phage HK629 and the other was an incomplete region closely related to Salmonella phage SPN9CC. The analysis using Prophinder detected not only the two prophage regions but also an additional region that was closely related to an enterobacteria lambda-like phage. All the prophage regions observed for E455 were also found in the draft genome sequence of E455-L. More importantly, E455-L was observed carrying an extra P2-like prophage, which was detected by all sequence analysis methods used (i.e. PHAST or Prophinder; raw or annotated). This P2-like prophage was identified to be 29.3–34.8 kb in size with 40–46 CDS, depending on the method of analysis being used (Table 1). The CDS include 31 phage structural genes, 6 phage regulatory genes, 1 putative reverse transcriptase and 8 genes of unknown functions (Supplementary Information 6, available online). The DNA sequence of the extra P2-like prophage in E455-L was identical to one of the two P2-like prophage regions observed for E1140. The phage–host junctions (attL, attR) of the P2-like prophage was predicted by using PHAST (Supplementary Information 6). The bacterial attachment sites (attB) in both E455-L and E1140 were identical. The attP sites of the P2-like prophage in E455-L was also identical to those in E1140. The P2-like prophage in E455-L was also inducible as indicated by the sequencing of the randomly amplified DNA of the phage particles induced from the strain (Supplementary Information 7). As indicated above, we isolated E455-L from a turbid plaque on the lawn of E455 treated with a phage-containing lysate of E1140. By analyzing the genomes of the three E. coli strains and the randomly amplified DNA for their induced phages, we ruled out the possibility of E455-L being a contaminant or a resistant mutant, and confirmed it as an isogenic lysogen of E455 that had been infected by a P2-like phage induced from E1140. Besides integration into host chromosomes, temperate phages may integrate into plasmids or exist as independent plasmids (Casjens, 2003). Presumably, such integration had not occurred for E455-L, since we did not observe differences in the number and sizes of plasmids extracted from E455 and E455-L (Supplementary Information 8, available online). Nonetheless, this needs to be confirmed by sequencing of the plasmids of the three E. coli strains. Growth and survivorship under different lab conditions All three strains exhibited typical log-linear growth when incubated in 1X TSB at 25, 30 or 37°C, which approximated the median and the highest temperatures of the surface water in Hong Kong (Hong Kong EPD which approximated the median and the highest temperatures of the surface water in Hong Kong (Hong Kong EPD 2016) and the body temperature of pigs, respectively. The specific growth rates of the three strains, as determined on the basis of the optical density of the batch cultures, increased significantly with incubation temperature (P ≤ 0.001) without significant differences between strains (P ≥ 0.058, Fig. 1). Figure 1. View largeDownload slide Growth of the three E. coli strains in TSB. Data shown are the maximum growth rates and the ln(OD600) value of the three E. coli strains when growing in TSB at (a) 25°C, (b) 30°C, and (c) 37°C. Error bars indicate ( ± 1) standard deviation (n = 3). Bars with a straight line above them are not significantly different from each other. Figure 1. View largeDownload slide Growth of the three E. coli strains in TSB. Data shown are the maximum growth rates and the ln(OD600) value of the three E. coli strains when growing in TSB at (a) 25°C, (b) 30°C, and (c) 37°C. Error bars indicate ( ± 1) standard deviation (n = 3). Bars with a straight line above them are not significantly different from each other. Contrasting with the continuous growth in nutrient-rich broth, the three strains underwent deactivation in seawater microcosms over 4 days of incubation at different rates as determined on the basis of colony-forming units (P < 0.05, Fig. 2c). The fecal strain E455 had a faster deactivation (4.38 ± 0.029 day−1) than the sediment strain E1140 (1.69 ± 0.02 day−1), while the lysogen E455-L had a deactivation rate (2.63 ± 0.08 day−1) in between those of E455 and E1140. The deactivation of E. coli in seawater is commonly attributed to the high salinity and low nutrient conditions as major stresses (Liu et al.2015). Figure 2. View largeDownload slide Survivorship of the three E. coli strains in microcosms. Data shown are the mean growth rates and the mean concentrations of the three E. coli strains when they were incubated in (a) sediment, (b) interstitial water, and (c) seawater at 30°C. Error bars indicate ( ± 1) standard deviation (n = 3). Bars with a straight line above them are not significantly different from each other. Bars assigned different letters are significantly different from each other. Figure 2. View largeDownload slide Survivorship of the three E. coli strains in microcosms. Data shown are the mean growth rates and the mean concentrations of the three E. coli strains when they were incubated in (a) sediment, (b) interstitial water, and (c) seawater at 30°C. Error bars indicate ( ± 1) standard deviation (n = 3). Bars with a straight line above them are not significantly different from each other. Bars assigned different letters are significantly different from each other. In contrast to the seawater microcosms, the three E. coli strains exhibited population growth in the first 24 h in the sediment microcosms followed by continuous deactivation over the next 18 weeks (Fig. 2a). The fecal strain deactivated the fastest (0.28 ± 0.01 day−1) while the sediment strain was the slowest (0.07 ± 0.01 day−1); the lysogen had an intermediate rate (0.14 ± 0.01 day−1). Nonetheless, all deactivation rates in sediment were significantly lower than those in the seawater microcosms. This phenomenon has been observed in many other studies and was attributed to the presence of higher amounts of organic nutrients in sediment and the sediment grains providing extensive surfaces for the formation of biofilms, which helped E. coli to survive (Gerba and McLeod 1976; Stephenson and Rychert 1982; Pommepuy et al.1992; Anderson, Whitlock and Harwood 2005; Thomason et al.2012). Prior to deactivation in the sediment microcosms, all three E. coli strains exhibited rapid growth within the first 24 h of incubation, with no significant differences in growth rates between them (maximum growth rate >4 day−1; P > 0.785). This observation suggested that the three E. coli strains were able to reproduce by utilizing the dissolved substrates of the sediment matrix as the sole source of nutrient (Garzio-Hadzick et al.2010). This is also supported by the growth observed in interstitial water microcosms, which contained the dissolved substrates of the sediment matrix (Fig. 2b). The maximum growth rates of the three strains in the interstitial water did not differ from each other (rate >1.5 day−1, P > 0.982). Nonetheless, the magnitude of growth was higher in the sediment than in the interstitial water, presumably due to the former providing surfaces for biofilm formation and containing a wider range of nutrients (Pommepuy et al.1992; Garzio-Hadzick et al.2010). The microcosm experiments were repeated and the results of the second trial are stated in Supplementary Information 9 (available online). The patterns of growth and deactivation of the three E. coli strains were consistently observed in the two trials. The resources and stresses that E. coli encounters in the animal body and the external environment are drastically different. Savageau (1983) hypothesized that E. coli was adapted to cycling between the two habitats through the fecal–oral route by having evolved a dual regulatory system to control gene expression under two different sets of environmental conditions. This is supported by the highly conserved ecophysiological traits observed for 25 E. coli strains of animal and environmental sources (Ihssen et al.2007). On the contrary, Whittam (1989) proposed that an E. coli population leaving the animal body was under strong selective pressure for traits suitable for survival in the external environment, and that some strains would be better adapted to the former and some to the latter. Ishii et al. (2006) further postulated that E. coli in the external environment could go through genetic changes, become fully adapted, and survive as an autochthonous population. The mechanism for E. coli to adapt to the contrasting habitats of animal bodies and external environment still remains unclear. Our results showed that isogenic lysogen E445-L had lower deactivation rates than its wild type E445 when exposed to seawater and sediment, although they did not differ in growth rates when exposed to nutrient-rich medium, sediment interstitial water or the sediment matrix during the first 24 h of incubation. Nonetheless, these results are in support of our hypothesis that lysogeny could be a mechanism to mediate the fitness of E. coli in the external environment. Phenotype profiling Compared to the relatively stable and resourceful animal guts, conditions in the natural environment are more dynamic and challenging. One of the major stresses that E. coli encounters in the external environment is nutrient depletion. The ability to exploit nutrients has been known to be a crucial factor for E. coli to grow and resist other stresses in the environment (Lopez-Torres, Prieto and Hazen 1988; Rozen and Belkin 2001; van Elsas et al.2011). In this study, we compared the three E. coli strains for their capacity in nutrient acquisition, storage and utilization. Glycogen is the primary energy storage for E. coli to withstand temporal nutritional deficiency (Boos and Shuman 1998; Wilson et al.2010). The accumulation of glycogen has been shown to support E. coli to persist and colonize the environment (Jones et al.2008). The amount of glycogen accumulated by E455, E455-L and E1140 on M9 agar (supplemented with 0.4% glucose), as indicated by color intensity of iodine stained on the bacterial colonies, was 124, 127 and 127 arbitrary units, respectively (Fig. 3a). There were no significant differences (P > 0.455) between the three strains over 2 days of incubation. Figure 3. View largeDownload slide Stress-resistant phenotypes of the three E. coli strains. (a) Glycogen accumulation on M9 agar, (b) swimming motility on motility agar, and (c) biofilm formation under 0.2X TSB and 1X TSB. Positive controls and negative controls were included in each experiment. Error bars indicate ( ± 1) standard deviation (n = 3). Bars assigned with different letters are significantly different from each other. Figure 3. View largeDownload slide Stress-resistant phenotypes of the three E. coli strains. (a) Glycogen accumulation on M9 agar, (b) swimming motility on motility agar, and (c) biofilm formation under 0.2X TSB and 1X TSB. Positive controls and negative controls were included in each experiment. Error bars indicate ( ± 1) standard deviation (n = 3). Bars assigned with different letters are significantly different from each other. Swimming motility and biofilm formation are related to the acquisition of nutrients in the environment. Among the three E. coli strains, E1140 formed significantly larger colonies (6.44 ± 0.81 cm2, P ≤ 0.001, Fig. 3b) on the nutrient-rich motility agar than the other two strains E455 (0.62 ± 0.16 cm2) and E455-L (0.66 ± 0.20 cm2), suggesting that only E1140 adopted a motile lifestyle when nutrients or energy were sufficient. E455 and E455-L did not differ significantly in swimming motility (P > 0.569). The results of the biofilm formation test also suggested that the fecal strains adopted a sessile lifestyle in nutrient-rich conditions (1X TSB, Fig. 3c). This is indicated by E455 and E455-L being able to form significantly stronger biofilms, which retained higher concentrations of crystal violet stains (A600 = 1.09 ± 0.17 and 1.06 ± 0.07, respectively) than did E1140 (0.37 ± 0.06) (P < 0.001). In contrast, none of the three strains were able to form detectable biofilm in relatively dilute medium (0.2X TSB, A600 < 0.3, P > 0.156, Fig. 3c). The results were consistent with other studies that E. coli motility and biofilm formation may interconvert in response to changes in nutrient availability (Thomason et al.2012). In the natural environment, the quality and quantity of carbon substrates vary substantially between habitats. The ability to utilize a wide range of substrates in an efficient way increases the chance for E. coli to persist in the environment and compete with the indigenous bacteria. In this study, the carbon utilization profile of the three E. coli strains was assessed in the OmniLog Phenotype Microarrays system. In this system, cellular respiration irreversibly reduces the soluble tetrazolium violet in the test medium to the insoluble formazan dye, which accumulates over time. Bacterial growth in response to a carbon substrate would result in exponential increase in formazan dye concentration whereas bacterial respiration without growth would result in only very slow accumulation of the dye. Therefore, the data of dye accumulation generated by the OmniLog Phenotype Microarrays are often used for the assessment of bacterial growth (Vorwerk et al.2014; Orro et al.2015). Among all sole carbon sources tested, 83 could not support the growth of any of the three E. coli strains while 32 supported the same level of growth in all three strains (P ≥ 0.184, Fig. 4). E455 and E455-L were able to grow in 25 carbons without differences between strains, but their growth differed significantly from that of E1140 (P ≤ 0.001, Fig. 4). Compared to E455, the utilization of 37 carbons (mostly carbohydrates) were up-regulated in E455-L and the utilization of 12 carbons were down-regulated (mostly carboxylic acids). In addition, the growth of E455-L in methyl pyruvate, propionic acid and l-glutamine reduced by 61–69% in comparison to that of E455. The growth of E1140 in these carbon substrates was also 25–33% lower than that of E455. Figure 4. View largeDownload slide Comparison of the three E. coli strains for differences in metabolic activity when utilizing alternative carbon sources (PM1 and PM2), under osmotic (PM9) or under pH (PM10) stresses. Reactions in PM1 and PM2 were categorized by the chemical nature of the carbon substrates. PM9 and PM10 were classified by the type of osmotic stress or pH stress acting on the E. coli. Data shown are the numbers of tested reactions that (i) all three strains showed no detectable growth, (ii) all three E. coli strains had detectable growth and without significant difference in their metabolic activity level, (iii) E455 and E455-L showed no significant difference, and (iv) E455 and E455-L showed significant differences. Figure 4. View largeDownload slide Comparison of the three E. coli strains for differences in metabolic activity when utilizing alternative carbon sources (PM1 and PM2), under osmotic (PM9) or under pH (PM10) stresses. Reactions in PM1 and PM2 were categorized by the chemical nature of the carbon substrates. PM9 and PM10 were classified by the type of osmotic stress or pH stress acting on the E. coli. Data shown are the numbers of tested reactions that (i) all three strains showed no detectable growth, (ii) all three E. coli strains had detectable growth and without significant difference in their metabolic activity level, (iii) E455 and E455-L showed no significant difference, and (iv) E455 and E455-L showed significant differences. The carbon utilization profile of E. coli is known to vary substantially between strains that occupy different niches in animal hosts. For example, the pathogenic E. coli O157:H7 strains exhibited carbon utilization patterns that were different from those of commensal isolates in 19 out of 95 carbon sources tested in the phenotype microarrays (Durso, Smith and Hutkins 2004). It was also shown that E. coli O157:H7 and the commensal strain EDL933 had different preferences for the variety of sugars present in mouse gut. These results are in support of the hypothesis that intraspecific diversification of carbon utilization profiles can reduce competition between strains that co-exist in the same habitat (Fabich et al.2008). In our case, the insertion of a prophage resulted in a shift in the recipient's (E455-L) carbon utilization profile in resemblance to that of the phage donor (E1140). This suggests that lysogeny is a possible mechanism for E. coli to acquire a new nutrient utilization profile that may help diversify their niches from the wild type strain. Apart from nutrient utilization, stress resistance is also an important fitness attribute for E. coli to survive in an external environment. The growth of the three E. coli strains under 192 osmotic and pH conditions were tested. All E. coli strains had growth in 122 test conditions. Among these conditions, E455 and E455-L had the same level of growth in 40 of them, and their growth rates were significantly different from that of E1140 (P ≤ 0.049). Compared to E455, E455-L had a larger magnitude of growth in 24 conditions (P ≤ 0.033) and smaller magnitude of growth in 6 conditions (P ≤ 0.033). In particular, E455-L grew faster than E455 in high salinity (NaCl 6%) and in the presence of six types of compatible solutes; γ-amino-N-butyric acid is the only osmolyte that appeared to suppress the growth of the former. These results suggest that osmotic regulation is a possible survival benefit conferred by the P2-like prophage. For the test of growth in different pH levels, the lysogen E455-L generally has similar or better growth than the wild type E455 in both acidic and alkaline conditions, except for the addition of two hydrophobic amino acids, norleucine and tryptophan, which resulted in the lysogen being less acid tolerant. Generally, the results obtained from the phenotypic microarrays indicated that E455 and E455-L had similar levels of growth under 80% of the conditions tested. In the remaining 20% of test conditions where the two strains had different growth rates, the response of lysogen E455-L appeared to be more similar to that of the prophage donor E1140, suggesting that the physiological changes between E445 and E455-L were brought about by the transferred P2-like prophage from E1140. Effects of lysogeny The effects of lysogeny on bacterial host physiology have been well studied through the elimination of prophage sequences from the wild type (Wang et al.2010; Yu et al.2015) or through the introduction of prophages to the hosts (Edlin, Lin and Bitner 1977). This is the first study that aims to investigate lysogeny as a possible mediator of E. coli fitness in external environments. We used a prophage donor isolated from marine sediment (E1140). Its induced P2-like prophage was able to infect a fecal strain E455 and generate an isogenic lysogen E455-L with some ecophysiological traits being more similar to the prophage donor than to the wild type parent strain. P2-like prophages have a wide distribution in E. coli genome and a broad host range within γ-proteobacteria. In the E. coli reference collection (ECOR) that contain strains from a wide range of animal hosts, 26% of the strains contain a P2-like prophage (Nilsson, Karlsson and Haggård-Ljungquist 2004). P2-like prophages were also found in many other species within the γ-proteobacteria including Haemophilus influenza (Esposito et al.1996), Pseudomonas aeuginosa (Nakayama et al.1999), Vibrio cholera (Nesper et al.1999) and Salmonella typhimurium (Mirold et al.2001). The broad host range of P2-like phages are attributed to their tail fibers being similar to those of several phage families (Mu, P1, λ, K3 and T2), making the P2-like phages able to alter host ranges under selective pressure (Haggård-Ljungquist, Halling and Calendar 1992). Currently, we are unable to determine whether E1140 is a long-term resident of the sediment or is derived from recent fecal pollution, and how E1140 has acquired the P2-prophage in the first place. Without any presumption of the actual ecological niche of E1140 or the history of the association between the P2-like prophage and E1140, the results in this study indicate that a prophage inserted into an E. coli host would inflict tractable changes in phenotypic traits, including enhanced survival in environmental matrices external to animal bodies. The traits were associated with both stress resistance and resources acquisition. In particular, the lysogen showed better growth than the wild type in a broad range of carbon substrates and under osmotic or pH stresses, while suppressed growth was observed only in a relatively small percentage of test conditions. These results altogether are in support of our hypothesis of lysogeny being a possible mechanism to mediate ecophysiological traits that have an effect on the fitness of E. coli in environments outside animal bodies. Bacteriophages can alter the host's physiology by bringing new functional genes to the bacterial genome (Kwon, Seong and Kim 2013; Wiles et al.2013), or they may carry genes that regulate the transcription of bacterial genes (Wang et al.2010). Some prophages can act as a regulatory switch to a bacterial operon when they insert into/excise from the bacterial chromosome (Feiner et al.2015). In this study, the molecular mechanisms causing the observed changes in the ecophysiological traits of the isogenic lyosgen are unclear. None of the 46 genes carried by the P2-like prophage (31 phage structural genes, 6 phage regulatory genes, 1 putative reverse transcriptase and 8 genes of unknown functions) have been reported to have effects on bacterial survival or nutrient utilization. In addition, the insertion site of the P2-like prophage was in a non-coding region of E455-L between a hypothetical protein and a putative protease with no operon being disturbed. It is also unlikely that the excision of the P2 prophage can act as a regulatory switch for gene functions either. There are insertion/deletion and substitution sites on the draft genome sequences. These could have effects on ecophysiological traits if they are not sequencing errors and they are located in a gene operon, which can only be determined by completing the genome sequences. Nonetheless, the relationship between E445 and E455-L as wild type and isogenic lysogen are well supported by a multitude of evidence (genome sequencing, PCR-DNA fingerprint of the genome DNA and phage DNA, and assay for superinfection immunity) and the enhanced survival of the lysogen E455-L in sediment and seawater was clear and repeatable. Therefore, the E. coli strains in this study can serve as a model to further investigate the molecular mechanisms and metabolic pathways that can affect the survival of E. coli in the natural environment. SUPPLEMENTARY DATA Supplementary data are available at FEMSEC online. FUNDING This study was supported by the General Research Fund (664012 and 661413), and Hong Kong University of Science and Technology School of Science allocation (FSGRF16SC06) to SCKL. Conflict of interest. None declared. REFERENCES Abràmoff MD, Magalhães PJ, Ram SJ. Image processing with ImageJ. Biophotonics Intern 2004; 11: 36– 42. Anderson KL, Whitlock JE, Harwood VJ. Persistence and differential survival of fecal indicator bacteria in subtropical waters and sediments. Appl Environ Microbiol 2005; 71: 3041– 8. Google Scholar CrossRef Search ADS PubMed Aziz RK, Bartels D, Best AA et al. The RAST server: rapid annotations using subsystems technology. BMC Genomics 2008; 9: 75. Google Scholar CrossRef Search ADS PubMed Badgley BD, Ferguson J, Heuvel AV et al. Multi-scale temporal and spatial variation in genotypic composition of Cladophora-borne Escherichia coli populations in Lake Michigan. Water Res 2010; 45: 721– 31. Google Scholar CrossRef Search ADS PubMed Barker C, Prüss BM, Matsumura P et al. Increased motility of Escherichia coli by insertion sequence element integration into the regulatory region of the flhD operon. J Bacteriol 2004; 186: 7529– 37. Google Scholar CrossRef Search ADS PubMed Bergthorsson U, Ochman H. Distribution of chromosome length variation in natural isolates of Escherichia coli. Mol Biol Evol 1998; 15: 6– 16. Google Scholar CrossRef Search ADS PubMed Berney M, Weilenmann HU, Ihssen J et al. Specific growth rate determines the sensitivity of Escherichia coli to thermal, UVA, and solar disinfection. Appl Environ Microbiol 2006; 72: 2586– 93. Google Scholar CrossRef Search ADS PubMed Boos W, Shuman H. Maltose/maltodextrin system of Escherichia coli: transport, metabolism, and regulation. Microbiol Mol Biol Rev 1998; 62: 204– 29. Google Scholar PubMed Brennan FP, O’Flaherty V, Kramers G et al. Long-term persistence and leaching of Escherichia coli in temperate maritime soils. Appl Environ Microbiol 2010; 76: 1449– 55. Google Scholar CrossRef Search ADS PubMed Burton E, Yakandawala N, LoVetri K et al. A microplate spectrofluorometric assay for bacterial biofilms. J Ind Microbiol Biotechnol 2007; 34: 1– 4. Google Scholar CrossRef Search ADS PubMed Byappanahalli MN, Whitman RL, Shively DA et al. Population structure, persistence, and seasonality of autochthonous Escherichia coli in temperate, coastal forest soil from a Great Lakes watershed. Environ Microbiol 2006; 8: 504– 13. Google Scholar CrossRef Search ADS PubMed Casjens S Prophages and bacterial genomics: what have we learned so far? Mol Microbiol 2003; 49: 277– 300. Google Scholar CrossRef Search ADS PubMed Chen Y, Golding I, Sawai S et al. Population fitness and the regulation of Escherichia coli genes by bacterial viruses. PLoS Biol 2005; 7: e229. Google Scholar CrossRef Search ADS Craig DL, Fallowfield HJ, Cromar NJ. Use of microcosms to determine persistence of Escherichia coli in recreational coastal water and sediment and validation with in situ measurements. J Appl Microbiol 2004; 96: 922– 30. Google Scholar CrossRef Search ADS PubMed Croxen MA, Finlay BB. Molecular mechanisms of Escherichia coli pathogenicity. Nat Rev Microbiol 2010; 8: 26– 38. Google Scholar CrossRef Search ADS PubMed Darling AE, Mau B, Blattner FR et al. Mauve: alignment of conserved genomic sequence with rearrangements. Genom Res 2004; 14: 1394– 403. Google Scholar CrossRef Search ADS Denisov G, Walenz B, Halpern AL et al. Consensus generation and variant detection by Celera Assembler. Bioinformatics 2008; 24: 1035– 40. Google Scholar CrossRef Search ADS PubMed Durso LM, Smith D, Hutkins RW. Measurements of fitness and Competition in Commensal Escherichia coli and E. coli O157:H7 Strains. Appl Environ Microbiol 2004; 70: 6466– 72. Google Scholar CrossRef Search ADS PubMed Edlin G, Lin L, Bitner R. Reproductive fitness of P1, P2, and Mu lysogens of Escherichia coli. J Virol 1977; 21: 560– 4. Google Scholar PubMed Esposito D, Fitzmaurice WP, Benjamin RC et al. The complete nucleotide sequence of bacteriophage HP1 DNA. Nucleic Acids Res 1996; 24: 2360– 8. Google Scholar CrossRef Search ADS PubMed Fabich AJ, Jones SA, Chowdhury FZ et al. Comparison of carbon nutrition for pathogenic and commensal Escherichia coli strains in the mouse intestine. Infect Immun 2008; 76: 1143– 52. Google Scholar CrossRef Search ADS PubMed Feiner R, Argov T, Rabinovich L et al. A new perspective on lysogeny: prophages as active regulatory switches of bacteria. Nat Rev 2015; 13: 641– 50. Garzio-Hadzick A, Shelton DR, Hill RL et al. Survival of manure-borne E. coli in streambed sediment: effects of temperature and sediment properties. Water Res 2010; 44: 2753– 62. Google Scholar CrossRef Search ADS PubMed Geeraerd AH, Valdramidis VP, Van Impe JR. GInaFiT, a freeware tool to assess non-log-linear microbial survivor curves. Int J Food Microbiol 2005; 102: 95– 105. Google Scholar CrossRef Search ADS PubMed Gerba CP, McLeod JS. Effect of sediments on the survival of Escherichia coli in marine waters. Appl Environ Microbiol 1976; 32: 114– 20. Google Scholar PubMed Goosney DL, Gruenheid S, Finlay BB. Gut feelings: enteropathogenic E. coli (EPEC) interactions with the host. Annu Rev Cell Dev Biol 2000; 16: 173– 89 Google Scholar CrossRef Search ADS PubMed Govons S, Vinopal R, Ingraham J et al. Isolation of mutants of Escherichia coli B altered in their ability to synthesize glycogen. J Bacteriol 1968; 97: 970– 2. Haggård-Ljungquist E, Halling C, Calendar R. DNA sequences of the tail fiber genes of bacteriophage P2: evidence for horizontal transfer of tail fiber genes among unrelated bacteriophages. J Bacteriol 1992; 174: 1462– 77. Google Scholar CrossRef Search ADS PubMed Hayashi T, Makino K, Ohnishi M. Complete genome sequence of enterohemorrhagic Escherichia coli O157: H7 and genomic comparison with a laboratory strain K-12. DNA Res 2001; 8: 11– 22. Google Scholar CrossRef Search ADS PubMed Hong Kong Environmental Protection Department. 2016. http://epic.epd.gov.hk/EPICRIVER/marine/?lang=zh (5 January 2018, date last accessed). Ihssen J, Grasselli E, Bassin C et al. Comparative genomic hybridization and physiological characterization of environmental isolates indicate that significant (eco-)physiological properties are highly conserved in the species Escherichia coli. Microbiology 2007; 153: 2052– 66. Google Scholar CrossRef Search ADS PubMed Ishii S, Ksoll WB, Hicks RE et al. Presence and growth of naturalized E. coli in temperate soils from Lake Superior. Appl Environ Microbiol 2006; 72: 612– 21. Google Scholar CrossRef Search ADS PubMed Ishii S, Sadowsky MJ. Escherichia coli in the environment: implications for water quality and human health. Microbes Environ 2008; 23: 101– 8. Google Scholar CrossRef Search ADS PubMed Ishii S, Yan T, Vu H et al. Factors controlling long-term survival and growth of naturalized Escherichia coli populations in temperate field soils. Microbes Environ 2010; 25: 8– 14. Google Scholar CrossRef Search ADS PubMed Johnson JR, Russo TA. Extraintestinal pathogenic Escherichia coli: “The other bad E. coli”. J Lab Clin Med 2002; 139: 155– 62. Google Scholar CrossRef Search ADS PubMed Jones SA, Jorgensen M, Chowdhury FZ et al. Glycogen and maltose utilization by Escherichia coli O157:H7 in the mouse intestine. Infect Immun 2008; 76: 2531– 40. Google Scholar CrossRef Search ADS PubMed Kaas RS, Friis C, Ussery DW et al. Estimating variation within the genes and inferring the phylogeny of 186 sequenced diverse Escherichia. coli genomes. BMC Genomics 2012; 13: 577. Google Scholar CrossRef Search ADS PubMed Kanehisa M, Goto S. KEGG: Kyoto encyclopedia of genes and genomes. Nucleic Acids Res 2000; 28: 27– 30. Google Scholar CrossRef Search ADS PubMed Krumsiek J, Arnold R, Rattei T. Gepard: a rapid and sensitive tool for creating dotplots on genome scale. Bioinformatics 2007; 23: 1026– 8. Google Scholar CrossRef Search ADS PubMed Kwon HJ, Seong WJ, Kim JH. Molecular prophage typing of avian pathogenic Escherichia coli. Vet Microbiol 2013; 162: 785– 92. Google Scholar CrossRef Search ADS PubMed Lai JYH, Zhang H, Chiang MHY et al. Draft genome sequences of three Escherichia coli strains investigated for the effects of lysogeny on niche diversification. Genome Announc 2014; 2: e00955– 14. Google Scholar PubMed Le Gall T, Clermont O, Gouriou S et al. Extraintestinal virulence is a coincidental by-product of commensalism in B2 phylogenetic group Escherichia coli strains. Mol Biol Evol 2007; 24: 2373– 84. Google Scholar CrossRef Search ADS PubMed Lima-Mendez G, Van Helden J, Toussaint A et al. Prophinder: A computational tool for prophage prediction in prokaryotic genomes. Bioinformatics 2008; 24: 863– 5. Google Scholar CrossRef Search ADS PubMed Liu R, Cheng KHF, Wong K et al. Differential utility of the Bacteroidales DNA and RNA markers in the tiered approach for microbial source tracking in subtropical seawater. Appl Microbiol Biotechnol 2015; 99: 5669– 81. Google Scholar CrossRef Search ADS PubMed Lopez-Torres AJ, Prieto L, Hazen TC. Comparison of the in situ survival and activity of Klebsilla pneumoniae and Escherichia coli in tropical marine environments. Microb Ecol 1988; 15: 41– 57. Google Scholar CrossRef Search ADS PubMed Luo C, Walk ST, Gordon DM et al. Genome sequencing of environmental Escherichia coli expands understanding of the ecology and speciation of the model bacterial species. Proc Natl Acad Sci U S A 2011; 108: 7200– 5. Google Scholar CrossRef Search ADS PubMed Mirold S, Rabsch W, Tschäpe H et al. Transfer of the Salmonella type III effector sopE between unrelated phage families. J Mol Biol 2001; 312: 7– 16. Google Scholar CrossRef Search ADS PubMed Moreira S, Brown A, Ha R et al. Persistence of Escherichia coli in freshwater periphyton: biofilm-forming capacity as a selective advantage. FEMS Microbiol Ecol 2012; 79: 608– 18. Google Scholar CrossRef Search ADS PubMed Nakayama K, Kanaya S, Ohnishi M et al. The complete nucleotide sequence of φCTX, a cytotoxin-converting phage of Pseudomonas aeruginosa: implications for phage evolution and horizontal transfer via bacteriophage. Mol Microbiol 1999; 31: 399– 419. Google Scholar CrossRef Search ADS PubMed Nesper J, Blass J, Fountoulakis M et al. Characterization of the major control region of Vibrio cholerae bacteriophage K139: immunity, exclusion, and integration. J Bacteriol 1999; 181: 2902– 13. Google Scholar PubMed Nilsson AS, Karlsson JL, Haggård-Ljungquist E. Site-specific recombination links the evolution of P2-like coliphages and pathogenic enterobacteria. Mol Biol Evol 2004; 21: 1– 13. Google Scholar CrossRef Search ADS PubMed Ohnishi M, Kurokawa K, Hayashi T. Diversification of Escherichia coli genomes: are bacteriophages the major contributors? Trends Microbiol 2001; 9: 481– 5. Google Scholar CrossRef Search ADS PubMed Orro A, Cappelletti M, D’Ursi P et al. Genome and phenotype microarray analyses of Rhodococcus sp. BCP1 and Rhodococcus opacus R7: genetic determinants and metabolic abilities with environmental relevance. PLoS One 2015; 10: e0139467. Google Scholar CrossRef Search ADS PubMed Pommepuy M, Guillaud JF, Dupray E et al. Enteric bacteria survival factors. Water Sci Technol 1992; 25: 93– 103. Rozen Y, Belkin S. Survival of enteric bacteria in seawater. FEMS Microbiol Rev 2001; 25: 513– 29. Google Scholar CrossRef Search ADS PubMed Savageau MA. Escherichia coli habitats, cell types, and molecular mechanisms of gene control. Am Nat 1983; 122: 732– 44. Google Scholar CrossRef Search ADS Stephenson GR, Rychert RC. Bottom sediment: a reservoir of Escherichia coli in rangeland streams. J Range Manage 1982; 35: 119– 23. Google Scholar CrossRef Search ADS Thomason MK, Fontaine F, De Lay N et al. A small RNA that regulates motility and biofilm formation in response to changes in nutrient availability in Escherichia coli. Mol Microbiol 2012; 84: 17– 35. Google Scholar CrossRef Search ADS PubMed Touchon M, Hoede C, Tenaillon O et al. Organised genome dynamics in the Escherichia coli species results in highly diverse adaptive paths. PLoS Genet 2009; 5: e1000344. Google Scholar CrossRef Search ADS PubMed Vaas LA, Sikorski J, Hofner B et al. opm: an R package for analyzing OmniLog® phenotype microarray data. Bioinformatics 2013; 29: 1823– 4. Google Scholar CrossRef Search ADS PubMed Van Elsas JD, Semenov AV, Costa R et al. Survival of Escherichia coli in the environment: fundamental and public health aspects. ISME J 2011; 5: 173– 83. Google Scholar CrossRef Search ADS PubMed Vehkala M, Shubin M, Connor TR et al. Novel R pipeline for analyzing Biolog phenotypic microarray data. PLoS One 2015; 10: e0118392. Google Scholar CrossRef Search ADS PubMed Veses-Garcia M, Liu X, Rigden DJ et al. Transcriptomic analysis of Shiga-toxigenic bacteriophage carriage reveals a profound regulatory effect on acid resistance in Escherichia coli. Appl Environ Microbiol 2015; 81: 8118– 25. Google Scholar CrossRef Search ADS PubMed Vorwerk H, Mohr J, Huber C et al. Utilization of host-derived cysteine-containing peptides overcomes the restricted sulphur metabolism of Campylobacter jejuni. Mol Microbiol 2014; 93: 1224– 45. Google Scholar PubMed Walk ST, Alm EW, Gordon DM et al. Cryptic lineages of the genus Escherichia. Appl Environ Microbiol 2009; 75: 6534– 44. Google Scholar CrossRef Search ADS PubMed Wang X, Kim Y, Ma Q et al. Cryptic prophages help bacteria cope with adverse environments. Nat Commun 2010; 1: 147. Google Scholar CrossRef Search ADS PubMed Whittam TS. Clonal dynamics of Escherichia coli in its natural habitat. Antonie Van Leeuwenhoek 1989; 55: 23– 32. Google Scholar CrossRef Search ADS PubMed Wick LM, Qi W, Lacher DW et al. Evolution of genomic content in the stepwise emergence of Escherichia coli O157: H7. J Bacteriol 2005; 187: 1783– 91. Google Scholar CrossRef Search ADS PubMed Wiles TJ, Norton JP, Smith SN et al. A phyletically rare gene promotes the niche-specific fitness of an E. coli pathogen during bacteremia. PLoS Pathog 2013; 9: e1003175. Google Scholar CrossRef Search ADS PubMed Wilson WA, Roach PJ, Montero M et al. Regulation of glycogen metabolism in yeast and bacteria. FEMS Microbiol Rev 2010; 34: 952– 85. Google Scholar CrossRef Search ADS PubMed Winfield MD, Groisman EA. Role of nonhost environments in the lifestyles of Salmonella and Escherichia coli. Appl Environ Microbiol 2003; 69: 3687– 94. Google Scholar CrossRef Search ADS PubMed Yu Z, Chen X, Shen Q et al. Filamentous phages prevalent in Pseudoalteromonas spp. confer properties advantageous to host survival in Arctic sea ice. ISME J 2015; 9: 871– 81. Google Scholar CrossRef Search ADS PubMed Zhou Y, Liang Y, Lynch KH et al. PHAST: a fast phage search tool. Nucleic Acids Res 2011; 39: 347– 52. Google Scholar CrossRef Search ADS PubMed © FEMS 2017. All rights reserved. For permissions, please e-mail: firstname.lastname@example.org
FEMS Microbiology Ecology – Oxford University Press
Published: Feb 1, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera