Insertion Sequence (IS)-Excision Enhancer (IEE)-Mediated IS Excision from the lacZ Gene Restores the Lactose Utilization Defect of Shiga Toxin-Producing Escherichia coli O121:H19 Strains and Is Responsible for Their Delayed Lactose Utilization Phenotype.

Lactose utilization is one of the general biochemical characteristics of Escherichia coli, and the lac operon is responsible for this phenotype, which can be detected on lactose-containing media, such as MacConkey agar, after 24 h of incubation. However, some Shiga toxin-producing E. coli (STEC) O121:H19 strains exhibit an unusual phenotype called delayed lactose utilization (DLU), in which lactose utilization can be detected after 48 h of cultivation but not after only 24 h of cultivation. Insertion of an insertion sequence (IS), IS600, into the lacZ gene appears to be responsible for the DLU phenotype, and exposure to lactose has been reported to be necessary to observe this phenotype, but the mechanism underlying these phenomena remains to be elucidated. Here, we performed detailed analyses of the lactose utilization abilities of a set of O121:H19 strains and their mutants and found that IS-excision enhancer (IEE)-mediated excision of IS600 reactivates the lacZ gene and that the selective proliferation of IS-cured subclones in lactose-supplemented culture medium is responsible for the expression of the DLU phenotype. In addition, we analyzed the patterns of IS insertion into the lacZ and iee genes in the global O121:H19 population and revealed that while there are O121:H19 strains or lineage/sublineages that contain the IS insertion into iee or intact lacZ and thus do not show the DLU phenotype, most currently circulating O121:H19 strains contain IS600-inserted lacZ and intact iee and thus exhibit this phenotype. IMPORTANCE Insertion sequences (ISs) can modulate gene expression by gene inactivation or activation. While phenotypic changes due to IS insertion/transposition are frequently observed, gene reactivation by precise or simple IS excision rarely occurs. In this study, we show that IS600 is excised from the lacZ gene by IS-excision enhancer (IEE) during the cultivation of Shiga toxin-producing Escherichia coli (STEC) O121:H19 strains that show an unusual phenotype called delayed lactose utilization (DLU). This excision rescued their lactose utilization defect, and the subsequent selective proliferation of IS-cured subclones in lactose-containing medium resulted in the expression of the DLU phenotype. As we also show that most currently circulating O121:H19 strains exhibit this phenotype, this study not only provides information helpful for the isolation and identification of O121:H19 STEC but also offers novel insights into the roles of IS and IEE in the generation of phenotypic variation in bacterial populations.

INTRODUCTION

Insertion sequences (ISs) are small transposable elements (typically 0.7 to 2.5 kb in size) that encode a transposase (TPase) and are usually flanked by terminal inverted repeats (IRs) (1). More than 4,000 ISs were deposited in the ISFinder database, and they are classified into 32 families based on several features, such as the sequences of their TPases and IRs (1, 2). IS transposition induces a variety of genome rearrangements, including deletion, inversion, and duplication (3, 4), and generates small structural polymorphisms (5), which have strong impacts on the genome diversification and evolution of bacteria (6).

Another important aspect of the impact of IS insertion/transposition on their bacterial hosts is the modulation of gene expression by gene inactivation or activation (7, 8). While phenotypic changes due to IS insertion are frequently observed in many bacterial species, there have been very few reports on gene reactivation by IS excision because precise or simple excision of IS elements rarely occurs in bacterial cells (9, 10). However, the IS-excision enhancer (IEE) identified in Shiga toxin-producing E. coli (STEC) strains promotes the excision of IS3 family members and generates various types of genomic alterations, including simple IS excision and deletion of IS-flanking regions (11). The iee gene is encoded by large integrative elements, SpLE1 in the O157:H7 strain Sakai and SpLE1-like elements in other STEC serotypes (O26:H11, O111:H8, O103:H2, O145:H28, and O121:H19) (11–13). IS3 family members, such as IS629, IS1203, and IS600, have been amplified in these STEC serotypes (12–16), and notable phenotypic and genotypic variations have been generated within these STECs by IS transposition and IEE-mediated IS excision (5, 15, 17, 18).

Lactose fermentation is one of the general biochemical characteristics of coliform bacteria belonging to four genera of Enterobacteriaceae (Citrobacter, Enterobacter, Escherichia, and Klebsiella) (19). E. coli utilizes lactose by hydrolyzing the β-1,4 glycosidic bond by β-galactosidase encoded by the lacZ gene, which forms the lac operon along with lacY and lacA encoding a permease and a transacetylase, respectively (20). While lactose utilization, which can be detected on lactose-containing agar plates, such as MacConkey agar (referred to as MAC), after 24 h of incubation is a trait often used to identify E. coli, it has recently been reported that some STEC O121 strains exhibit an unusual phenotype called delayed lactose utilization (DLU) (21). In DLU strains, lactose utilization was observed on MAC after 48 h of incubation but not after only 24 h of incubation, and exposure to lactose was necessary to observe this phenotype. As a copy of IS600 was inserted into lacZ in the DLU strains, this IS insertion appears to be related to the DLU phenotype, but the mechanism underlying this phenomenon has not been previously elucidated.

In this study, we performed detailed analyses of the lactose utilization ability of a set of O121:H19 strains and their mutants under several culture conditions and revealed that IEE-mediated IS600 excision reactivates the lacZ gene and that the selective proliferation of subclones that contain the IS-cured lacZ gene in lactose-containing medium is responsible for the expression of the DLU phenotype. In addition, we analyzed the patterns of IS insertion into lacZ and iee in the global O121H19 population and showed that while most currently circulating O121:H19 clones exhibit the DLU phenotype, some O121:H19 strains or sublineages contain IS insertion into iee and thus do not show this phenotype.

RESULTS

IS600 excision from the lacZ gene during cultivation.

As the IEE promotes the excision of IS elements belonging to the IS3 family (11) and O121:H19 strains contain an SpLE1-like element that encodes the iee gene (13), we hypothesized that the DLU phenotype is caused by IEE-mediated IS600 excision from lacZ. By reinspecting the genome sequences of O121:H19 strains (13) available in our laboratory, we identified three types of strains that differ in the intactness of lacZ and iee: (i) IS600-inserted lacZ (referred to as lacZIS) and wild-type iee (ieeWT), (ii) lacZIS and IS1203-inserted iee (ieeIS), and (iii) wild-type lacZ (lacZWT) and ieeWT (Fig. 1). We selected three representative strains, one of each type (the genome sequences of these strains have been finished [13]) (Table 1) and cultured them for 40 h on MAC. Strains E15042 and SE14002 (types ii and iii, respectively) clearly showed a negative and positive lactose-fermentation phenotype, respectively. However, in strain 51104 (type i), red microcolonies were formed in translucent colonies, indicating the emergence of lactose-fermenting subclones during colony growth (Fig. 1). By subculturing the colony of strain 51104 onto MAC, we obtained both translucent and red colonies after 16 h of incubation. PCR analysis of the genomic DNA extracted from these colonies revealed that IS600 was not present in the lacZ gene in red colonies but was present in that of translucent colonies (Fig. 1 and Fig. S1). As IS600 insertion occurred very close to the 3′ end of lacZ (Fig. S2A), it is not clear whether the IS insertion inactivated β-galactosidase. We therefore examined the β-galactosidase activity of each colony by the 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-gal) assay and found that lacZIS colonies did not show β-galactosidase activity (Fig. 1 and Fig. S2B). These results indicated that in the ieeWT strain, the excision of IS600 from lacZ occurred during cultivation on MAC, resulting in the emergence of β-galactosidase (lacZ)-reactivated subclones.

FIG 1

IS insertion into the lacZ and iee genes and lactose utilization abilities of three O121:H19 strains. In the left part, IS insertions into lacZ and iee in three O121:H19 strains are shown. These strains belong to different lineages or sublineages, as shown in parentheses (see Fig. 5A and the Results section for details). In the middle part, colonies of the O121:H19 strains cultured for 40 h on MAC and colonies reisolated from their 16 h subculture on MAC are shown. In the right part, the results of PCR analysis to detect the IS-inserted lacZ gene (lacZIS) and the β-galactosidase activity of each strain are shown (Fig. S1 and S2 for more details of the results).

TABLE 1

Strains and plasmids used for the analysis of lactose utilization

Strain or plasmid	Descriptiona	Source or reference(Accession No.)
Strains
51104	O121:H19; lacZIS, ieeWT	(13) (AP024471-AP024472)
51104Δiee	51104 derivative; lacZIS, iee::CmR	This study
E15042	O121:H19; lacZIS, ieeIS	(13) (AP024478-AP024479)
SE14002	O121:H19; lacZWT, ieeWT	(13) (AP024473-AP024474)
K-12	Wild type; lacZWT, ieeWT	(38) (NC_000913)
K-12_lacZIS	K-12 derivative; lacZIS	This study
K-12_lacZIS:pBR	K-12_lacZIS derivative carrying pBR322; ApR, TcR	This study
K-12_lacZIS:piee	K-12_lacZIS derivative carrying pBR-iee; TcR	This study
Plasmids
pBR322	Cloning vector; ApR, TcR	TaKaRa (J01749)
pBR-iee	pBR322 derivative with ECs1305 (iee) and its flanking region; TcR	This study

lacZWT, wild-type lacZ; lacZIS, IS600-inserted lacZ; ieeWT, wild-type iee, ieeIS, IS1203-inserted iee; CmR, chloramphenicol resistance; ApR, ampicillin resistance; TcR, tetracycline resistance.

FIG 5

Phylogeny of STEC O121:H19 and variable IS insertion into the lacZ and iee genes. (A) Phylogenetic relationship of the 442 O121:H19 strains analyzed. This unrooted ML tree was taken from our previous paper (13) and modified. Three sublineages in the L1 lineage defined in this study (L1.1, L1.2, and L1.3) are indicated, and the numbers of strains belonging to each sublineage are shown in parentheses. The numbers of strains and the proportions of wild-type or IS-inserted lacZ and iee genes in the four lineages are presented in the inset. In the genome sequence of an L1.1 strain (FWSEC0155), the iee gene was split, but IS insertion was not detected; therefore, this strain was excluded from calculation. (B) IS insertion into the lacZ and iee genes in lineage L1. An ML tree of 418 L1 strains was constructed based on the recombination-free SNPs (7,591 sites) identified on the chromosomal backbone sequence (3,704,750 bp) using an L3 strain (SE14002) as the outgroup. IS elements inserted into lacZ and iee in each strain were mapped on the tree. As shown in the lower panel, where IS positions in lacZ and iee are indicated by colored triangles, the identified IS elements were distinguished based on their insertion position.

Involvement of IEE in IS600 excision from lacZ.

The finding that red microcolonies representing lacZ-reactivated subclones were not formed on MAC in strain E15042 (type ii; carrying ieeIS) (Fig. 1) suggested that IEE is responsible for the excision of IS600 from lacZ. To verify this hypothesis, we constructed an iee deletion mutant of strain 51104 (51104Δiee) and two K-12 derivatives that carried the same lacZIS gene as that in strain 51104 and either a plasmid encoding the iee gene (K-12_lacZIS:piee) or an empty plasmid vector (K-12_lacZIS:pBR) (Table 1), and we examined the emergence of subclones carrying the IS600-excised lacZ gene after 40 h of cultivation on MAC (Fig. 2). Red (lactose-fermenting) colonies emerged from the wild-type 51104 strain and K-12_lacZIS:piee but not from 51104Δiee and K-12_lacZIS:pBR. In the two iee-negative mutants, lactose-fermenting colonies were not detected even in the spots of the bacterial suspension (1.3 × 107 CFU/spot). These results indicated that the excision of IS600 from lacZ is mediated by IEE.

FIG 2

Involvement of IEE in the emergence of lactose-fermenting subclones during cultivation and the effect of lactose. The proportion of lactose-fermenting subclones in the colonies of each strain cultured on MacConkey agar base plates supplemented with or without 1% lactose for 40 h at 37°C was determined by suspending 5 colonies in LB, inoculating the suspensions on MacConkey agar base supplemented with 1% lactose (MAC plate), and counting red colonies after 16 h of incubation at 37°C. For spot analyses, cell suspensions (approximately 1.3 × 107 CFU) were spotted onto MAC plates and grown for 16 h at 37°C, and representative patterns are shown for each strain and culture condition.

Increase in the subpopulation carrying IS600-excised lacZ in the stationary phase during cultivation in lactose-supplemented medium.

When the wild-type 51104 strain and the K-12_lacZIS:piee strain were cultured for 40 h on nonlactose-supplemented MAC (MacConkey agar base) plates, no lactose-fermenting colonies were obtained, although small numbers of red colonies were detected in the spot of the 51104 suspension (Fig. 2). IEE-mediated IS excision requires an active IS TPase (11); however, the TPase gene in the IS600 copy inserted into lacZ is transcribed in the direction opposite to that of the lac operon (Fig. 1), and thus, it is likely that the increased IS600 TPase expression by the induction of the lac operon is not involved in the enhancement of IS600 excision. To verify this hypothesis, we examined the effect of isopropyl-β-d-thiogalactopyranoside (IPTG) supplementation on the emergence of subclones carrying IS600-excised lacZ (Fig. S3). K-12_lacZIS:piee cultured in lysogeny broth (LB) supplemented with IPTG yielded a small number of lactose-fermenting subclones, but the numbers of such subclones were similar to or even lower than those in LB medium not supplemented with IPTG. This result suggests that the induction of the lac operon has no detectable impacts on the IS600 excision frequency.

We next analyzed the growth of the K-12_lacZIS:piee strain in LB supplemented with either lactose or maltose (1% each; wt/vol), the latter of which is also utilized by E. coli as a carbon source (22) (dotted lines in Fig. 3A). The addition of maltose increased the bacterial density in the stationary phase. In contrast, while the strain cultured in the presence of lactose grew similarly to the control culture (supplemented with neither lactose nor maltose) until 6 h after inoculation, it exhibited further growth between 9 h and 18 h after inoculation. This result suggested the possibility that this second growth phase represents the growth of subclones carrying IS600-excised lacZ by utilizing lactose. To verify this hypothesis, we monitored the temporal change in the proportion of these subclones in the cultures grown in each medium by determining the copy number of lacZWT relative to that of the lacY gene (solid lines in Fig. 3A). The relative copy number of lacZWT was maintained at a low level in the strain cultured in LB and maltose-supplemented LB (2~3 × 10−3 copies) throughout the 24-h cultivation. In contrast, in the strain cultured in lactose-supplemented LB, the relative copy number of lacZWT started to increase at 6 h and reached approximately 3 × 10−1 copies at 18 h. When various amounts of glucose were added to the medium, the proportion of lacZWT-carrying subclones decreased in a glucose concentration-dependent manner (Fig. 3B), indicating that lactose utilization was inhibited by carbon catabolite repression (23).

FIG 3

The emergence of lactose-fermenting subclones during the culture of K-12_lacZIS:piee in LB and the chemically defined minimal medium supplemented with various carbon sources. (A) Growth of K-12_lacZIS:piee cultured in LB or LB supplemented with lactose (1%) or maltose (1%) and the changes in the copy number of IS-excised lacZ during cultivation are shown by dotted lines. The relative copy numbers of IS-excised lacZ in the total cellular DNA were determined at each time point by calculating the ratio of the copy number of IS-excised lacZ relative to the copy number of the lacY gene and are shown by solid lines. (B) The proportions of lactose-fermenting subclones in the 18-h cultures of K-12_lacZIS:piee grown in LB supplemented with 1% lactose and different concentrations of glucose. The proportions were calculated by dividing the number of red colonies by the number of total colonies on MAC plates. The mean values with standard deviations of biological triplicates are shown. (C and D) Growth curves of K-12_lacZIS:piee cultured in MM and MM supplemented with 1% lactose or 1% maltose (C) and the changes in the copy number of IS-excised lacZ during cultivation (D). The relative copy numbers of IS-excised lacZ in the total cellular DNA were determined at 18 h and 36 h (indicated by arrowheads in Fig. 3C) as described for panel A. (E) The proportions of lactose-fermenting subclones in the 18-h cultures of K-12_lacZIS:piee grown in MM supplemented with 1% lactose and different concentrations of glucose. The proportions were calculated as described for panel B. All OD600 values and relative copy numbers in this figure are presented as the mean value with standard deviations of biological triplicates and are shown on a logarithmic scale.

We further performed similar analyses of K-12_lacZIS:piee using a chemically defined minimal medium (MM, see Materials and Methods for its composition) (Fig. 3C to E). These analyses also clearly showed the maltose utilization and delayed lactose utilization of the strain (Fig. 3C), the increase in the relative copy number of lacZWT by delayed lactose utilization (Fig. 3D), and the inhibition of lactose utilization by carbon catabolite repression (Fig. 3E).

In summary, we concluded that lactose does not affect IEE-mediated IS600 excision, but the presence of lactose in culture medium promotes the selective proliferation of subclones carrying the IS600-excised lacZ gene, resulting in the formation of lactose-utilizing microcolonies after extended incubation, as illustrated in Fig. 4, which was recognized as the DLU phenotype.

FIG 4

Schematic representation of the growth of E. coli harboring the lacZIS and ieeWT genes during cultivation in a medium supplemented with lactose and glucose.

Variable IS insertion into the lacZ and iee genes in the O121:H19 lineage.

As our preliminary inspection of O121:H19 genomes suggested that there is some variation in IS insertion into the lacZ and iee genes among O121:H19 strains, the DLU phenotype appears to be a strain- or lineage-associated phenotype. To understand the phylogenetic background underlying the appearance of the DLU phenotype, we analyzed the variation in IS insertion into the lacZ and iee genes in the 442 O121:H19 strains (all were sequence type [ST] 655 or single locus variants of ST655) used in our previous phylogenetic analysis of the global O121:H19 population, in which O121:H19 strains were divided into four distinct lineages (L1-L4) (Data set S1) (13). This analysis revealed that IS insertion into the two genes occurred only in the major lineage, L1 (Fig. 5A). Although IS600 insertion into lacZ was found in 81% of the L1 strains (338/418), IS insertion into iee (or the absence of iee) was detected in only 11% (48/418) of the L1 strains, and the remaining strains contained an intact iee. Therefore, most L1 strains exhibit the DLU phenotype upon extended cultivation in a lactose-supplemented medium, such as MAC.

We performed a more detailed analysis of the IS insertion into lacZ and iee in the L1 lineage. This lineage was divided into three sublineages (referred to as L1.1, L1.2, and L1.3) (Fig. 5), and of the three strains used for the above-mentioned analyses, strains 51104 and E15042 belonged to L1.1 and L1.3, respectively (strain SE14002 belonged to L3). While both lacZ and iee were intact in all L1.2 strains (n = 3), IS600 insertion into lacZ was detected in most L1.1 and L1.3 strains (305/381 and 33/34, respectively) at the same position (Fig. 5B). This indicates that IS600 insertion into lacZ occurred in the common ancestor of L1, and the IS600 copy was deleted in sublineage L1.2 by the function of IEE. Although iee was present in all L1.1 and L1.3 strains except for five (3 in L1.1 and 2 in L1.3, in which the iee-encoding SpLE1-like integrative element appears to have been deleted), the pattern of IS insertion into iee was very different from that in lacZ (Fig. 5B). In L1.1, IS insertion into iee was detected sporadically (in only 10 strains), and the types and positions of IS varied between strains (black triangles in Fig. 5B), suggesting the independent insertion of these ISs in these strains. In contrast, the IS1203 insertion found in strain E15042 (indicated as ieeIS-2 in Fig. 5B) was detected in most iee-positive L1.3 strains (31/32) (Fig. 5B and Data set S1). Interestingly, one or two additional IS insertions into iee were found in many of the ieeIS-2-carrying strains (26/32, at five insertion sites; all were IS600 insertions), although the biological significance of this IS accumulation is currently unknown. IS600 insertion was also found in an ieeIS-2-negative L1.3 strain; thus, iee was inactivated by IS insertion in all L1.3 strains analyzed here. These results suggest that among the L1 strains, most L1.1 strains exhibit the DLU phenotype, but the L1.2 (lacZ is intact) and L1.3 strains (in which iee has been inactivated) do not exhibit this phenotype.

DISCUSSION

This study revealed that the IEE-mediated excision of IS600 from lacZ is responsible for the DLU phenotype observed in STEC O121:H19. Although IS600 excision occurs at a low frequency, subclones carrying IS600-excised lacZ can selectively proliferate by using lactose during extended incubation, which explains why this phenotype can be observed only on lactose-containing agar plates. As the emergence of lactose-negative colonies from IS-cured subclones was not observed even after repeated subculturing on MAC agar plates, reversion to the negative phenotype is a very rare event at least in vitro. Although inactivation of the lacZ gene and switching lactose fermentation phenotype may confer some metabolic advantages to the O121:H19 strains carrying the IS600-inserted lacZ gene in some environments where carbon sources other than lactose are enriched, actual metabolic advantages by this mechanism is unknown.

O121:H19 STEC is one of the major STECs along with the O157, O26, O103, O111, O145, and O45 STECs (24) and is causing many outbreaks and sporadic cases of infection in the USA (25), Canada (26), and Japan (27). Of the four lineages of O121:H19, the major lineage (L1) is a globally circulating strain (13). The common ancestor of L1 strains acquired several virulence gene-encoding mobile genetic elements, such as the locus of enterocyte effacement (LEE), a virulence plasmid, and the Stx2 phage (13). Our current analysis revealed that IS600 insertion into lacZ also occurred in the common ancestor of L1 and that the iee gene is intact in most strains belonging to the L1.1 lineage, which is the major sublineage of L1 (Fig. 5B). Thus, the majority of O121:H19 strains isolated worldwide exhibit the DLU phenotype, although we should remember that the L2, L3, and L4 strains and the L1.2 strains show typical lactose fermentation and that the L1.3 strains and a few L1.1 strains are negative for lactose fermentation. These data are helpful in the screening, isolation, and identification of O121:H19 STEC.

We found that the cultivation of bacteria carrying both iee and IS3 family members easily generates phenotypic changes in the population. In the O121:H19 strain 51104, a total of 27 copies of IS600 (22 on the chromosome and five on the plasmid) as well as multiple intact copies of other IS3 family members, such as IS629 (10 copies) and ISCfr6 (six copies) were present (13). Thus, IEE-mediated phenotype changes other than that of lactose fermentation could take place in strain 51104 and probably in other O121:H19 strains containing the intact iee gene. Major STECs contain iee and multiple copies of IS3 family members, such as IS629 and IS600 (which are often extensively amplified), as does ETEC O139 isolated from swine (12–15, 28). Therefore, gene inactivation/reactivation by IS elements and IEE can occur also in these STEC and ETEC strains, which may cause changes in various phenotypes in nature and during cultivation in the laboratory. Moreover, as many IEE homologs have been identified in a broad range of bacterial species (11), although their functions have yet to be analyzed, IEE-mediated IS excision may also play an important role in phenotypic changes in other bacterial species.

MATERIALS AND METHODS

Bacterial strains and media.

The E. coli strains and their mutants used in this study are listed in Table 1. Whole-genome sequences of the 442 O121:H19 strains (Data set S1) previously used in our phylogenetic analysis (13) were used for the analysis of IS insertion into the lacZ and iee genes.

Bacteria were grown in the following media: LB (1% [wt/vol] Bacto Tryptone, Gibco; 0.5% [wt/vol] Bacto Yeast Extract, Becton, Dickinson [BD]; 1% [wt/vol] sodium chloride, nacalai tesque), LB agar (LB containing 1.5% [wt/vol] Bacto Agar, BD), MAC (Difco MacConkey agar base, BD; 1% [wt/vol] lactose monohydrate, Wako), MAC not supplemented with lactose (Difco MacConkey agar base), Pearlcore MAC (Pearlcore MacConkey agar, Eiken Chemical Co.), MM (Difco M9 Minimal Salts, BD; 2 mM magnesium sulfate heptahydrate, Wako; 0.1% D-[+]-glucose, nacalai tesque), and MM agar (MM containing 1.5% [wt/vol] Bacto Agar). The growth media were supplemented with regents and antibiotics when necessary at the following concentrations: L(+)-arabinose (Wako), 1 mM; IPTG (Wako), 0.3 mM or 30 mM; X-gal (TaKaRa), 40 μg/mL; sucrose (nacalai tesque), 10% (wt/vol); D-(+)-glucose, 0.1%, 0.2%, or 0.4% (wt/vol); lactose monohydrate, 1.0% (wt/vol); D(+)-maltose monohydrate (Wako), 1.0% (wt/vol); chloramphenicol (Wako), 20 μg/mL; ampicillin (Sigma), 50 μg/mL; tetracycline (nacalai tesque), 10 μg/mL.

Construction of mutant strain.

Mutant strains were generated as follows using the plasmid vectors and primers listed in Tables 2 and 3, respectively. To disrupt the iee gene in the O121:H19 strain 51104, we first introduced the Red recombinase-encoding pKD46 (29) into the strain. A DNA fragment containing the chloramphenicol resistance (CmR) cassette and terminal 55-nt extensions homologous to the iee-flanking region was generated by the 2-step tailed-PCR method using the two sets of primers (iee-H1-F1/R1 and iee-H1-F2/R2) and pKD3 as a template. The 1.1-kbp PCR product was purified, treated with DpnI, and transformed into 51104 carrying pKD46, in which arabinose-inducible Red recombinase was expressed. Disruption of the iee gene in CmR transformants was confirmed by colony PCR using EmeraldAmp MAX PCR Master Mix (TaKaRa) and specific primers (Viee-F/R).

TABLE 2

Primers used to generate mutant strains

Name	Sequence (5′–3′)	Regions/positions of boldfaced sequences
iee gene disruption primers
iee-H1-F1	AACCCGGGGAGCTCAAATTATTTA-AAAAGTGTAGGCTGGAGCTGCTTC	pKD3
iee-H1-R1	TCAGTGGATCGGATACAGGTAAC-GAATATGGGAATTAGCCATGGTCC	pKD3
iee-H1-F2	TATACTGGGACTCTTTGTTGCCC-GAACAACCCGGGGAGCTCAAAT	5′-terminal region of the PCR product amplified with the iee-H1-F1 and -R1 primers
iee-H1-R2	AGGGGACCCGAATTTCTGCCGTG-GTGCCTCAGTGGATCGGATACA	3′-terminal region of the PCR product amplified with the iee-H1-F1 and -R1 primers
Primers to verify disruptions
Viee-F	GTTCTTACTGCCGGTAGCCATT
Viee-R	CCTCATTACAGTAATGCGGT
Primers to construct the lacZIS-pABB-CRS2 plasmid
lacZIS-UF	AGAGCTCGGATCCACTAGCGTAG-TGCAACCGAACGCGA	Terminus of the XbaI and SpeI-digested pABB-CRS2 plasmid
lacZIS-UR	GGTAATGGTAGCGACCGGCGCTCA	5′-terminal region of the PCR product amplified with the lacZIS-F and -R primers
lacZIS-F	GGTCGCTACCATTACCTGAG	3′-terminal region of the PCR product amplified with the lacZIS-UF and -UR primers
lacZIS-R	CACCAGACCAACTGGTTGAG	5′-terminal region of the PCR product amplified with the lacZIS-DF and -DR primers
lacZIS-DF	AAAGGATCGATCCTCTAGACAG-ACATAATAGTGCCAGC	3′-terminal region of the PCR product amplified with the lacZIS-F and -R primers
lacZIS-DR	ACCAGTTGGTCTGGTGTCAA	Terminus of the XbaI and SpeI-digested pABB-CRS2 plasmid
Primers to verify the insertion of three amplicons into the XbaI and SpeI-digested pABB-CRS2 plasmid
VlacZIS-pABB-CRS2-F	GTTGCATGGGCATAAAGTTG
VlacZIS-pABB-CRS2-R	CGGCTGACATGGGAATTCTA
Primers to verify integration of the lacZIS-pABB-CRS2 plasmid
Vint-lacZIS-pABB-CRS2-F1	GGGAAGTAGGCTCCCATGA
Vint-lacZIS-pABB-CRS2-R1	CCTCTGGATGTCGCTCCACA
Vint-lacZIS-pABB-CRS2-F2	CAACTGATGGAAACCAGCC
Vint-lacZIS-pABB-CRS2-R2	CTGAGGTGGCGAACGATGAG
Primers to verify the insertion of IS600 into the lacZ
VlacZIS-F	GGGAAGTAGGCTCCCATGA
VlacZIS-R	CAACTGATGGAAACCAGCC

TABLE 3

Plasmids used to generate mutant strains

Plasmids	Descriptiona	Source or reference
pKD46	ApR; oriR101-derived Red recombinase expression plasmid	29
pKD3	CmR; cloning vector	29
pABB-CRS2	ApR, CmR; R6K-derived suicide vector	30
lacZIS-pABB-CRS2	ApR; 51104 IS600-inserted lacZ cloned into pABB-CRS2	This study

ApR, ampicillin resistance; CmR, chloramphenicol resistance.

To generate a K-12 mutant that carried the same lacZIS gene as strain 51104 (K-12_lacZIS), we constructed a pABB-CRS2-based suicide vector (30), as shown in Fig. S4. Flanking regions of IS600 in the lacZ were amplified using two sets of primers (lacZIS-UF/UR and lacZIS-DF/DR) that included a 3-nt sequence of the target sequence for IS600 or an 18-nt sequence homologous to pABB-CRS2 at each 5′ end and the genomic DNA of K-12 as a template to obtain two amplicons. The IS600 sequence with a 13-bp sequence homologous to these two amplicons at each terminal end was also amplified using primers lacZIS-F/R and the genomic DNA of strain 51104 as a template. These three amplicons were purified by gel extraction and cloned into XbaI- and SpeI-digested pABB-CRS2 by the Red/ET recombination-based seamless DNA cloning method (SLiCE method) (31). The recombinant product was introduced into One Shot PIR1 Chemically Competent E. coli (Invitrogen) by transformation. Clones were selected on LB agar containing ampicillin (Ap plate). The insertion of the amplicons into pABB-CRS2 was confirmed by colony PCR using specific primers (VlacZIS-pABB-CRS2-F/R). The purified plasmid (lacZIS-pABB-CRS2) was transformed into K-12 using an electroporator (Gene Pulser II; Bio–Rad), followed by selection on Ap plates. Chromosome integration of lacZIS-pABB-CRS2 by a single crossover was confirmed by PCR using two sets of primers (Vint-lacZIS-pABB-CRS2-F1/R1 and Vint-lacZIS-pABB-CRS2-F2/R2), and total cellular DNA was extracted from isolated colonies using the alkaline-boiling method. To allow the excision of the sacB-containing vector, the transformants were grown in LB containing sucrose without NaCl for 6 h at 30°C with shaking and spread on LB agar plates containing sucrose without NaCl at 30°C. Finally, the presence of the lacZIS gene in each sucrose-resistant and Ap-sensitive colony was confirmed by PCR using specific primers (VlacZIS-F/R).

We constructed pBR-iee using pBR322 (Accession No. J01749; TaKaRa) by a similar strategy to that described for lacZIS-pABB-CRS2 using the primers listed in Table 4. The iee-encoding region was amplified using the genomic DNA of the O157:H7 strain Sakai (32) as a template and primers that included 15-nt sequences homologous to pBR322 at each 5′ end (ECs1305-pBR-F/R). The amplicon was purified by gel extraction and cloned into EcoRI/PstI-digested pBR322 by the SLiCE method. The recombinant product was introduced into NEB 5α (New England Biolabs) by transformation. Clones were selected on LB agar containing tetracycline. Insertion of the amplicon into pBR322 was confirmed by colony PCR using specific primers (VpBR-iee-F/R). pBR322 and pBR-iee were introduced into K-12_lacZIS using an electroporator (Gene Pulser II), and plasmid-carrying clones were selected based on tetracycline resistance.

TABLE 4

Primers used to construct pBR-iee

Name	Sequence (5′–3′)
Primers to construct pBR-iee plasmid
ECs1305-pBR-F	TCAAACATGAGAATTCTGAATTACAGCTGTATTAGG
ECs1305-pBR-R	GTTGCCATTGCTGCATTTTTGATGACGTTCTCACTG
Primers to verify the insertion of iee and its flanking region into pBR322 plasmid
VpBR-iee-F	CGTGTTATAGGTTGTGCGGAGA
VpBR-iee-R	TTTGCAAGCAGCAGATTACG

The boldfaced sequence in ECs1305-pBR-F and -R correspond to each terminal of the EcoRI and PstI-digested pBR322 plasmid.

Analysis of lactose fermentation, bacterial growth, and the copy number of lacZ gene.

Seven experiments outlined in Fig. S5 were performed in this study. In all experiments, tetracycline (10 μg/mL at the final concentration) was added to the culture medium for the maintenance of pBR322 and pBR-iee in K-12_lacZIS.

Experiment 1. Three O121:H19 strains (51104, E15042, and SE14002) were streaked onto a MacConkey agar base supplemented with lactose (MAC plates) and cultured for 40 h at 37°C. Then, single colonies were subcultured on MAC for 16 h at 37°C. Red or white colonies (lactose fermenting or not, respectively) were suspended in LB, and total cellular DNA was extracted from the suspension using the alkaline-boiling method and used for PCR analysis with the primers shown in Fig. S5. The remaining suspension was diluted to 1 × 109 CFU/mL (1 OD600), and the diluent (5 μL) was spotted onto LB agar supplemented with 0.3 mM IPTG (Wako) and X-gal (TaKaRa) and cultured for 16 h at 37°C. The β-galactosidase activity in each suspension was judged by the hydrolysis of X-gal (blue colored spot). Two lacY-deficient K-12 derivatives carrying the wild-type or degraded lacZ gene (HB101 or JM109, respectively; TaKaRa) were used as controls.

Experiment 2. Four strains (O121:H19 51104, 51104Δiee, K-12_lacZIS:pBR, and K-12_lacZS:piee) were streaked onto MAC and MacConkey agar base plates and cultured for 40 h at 37°C. Five colonies on each plate were randomly selected and suspended in LB. Clones in these suspensions were examined by the following procedures:

The suspensions were diluted to 1~2 × 103 CFU/mL, and each diluent (100 μL) was inoculated onto Pearlcore MAC plates. After incubation for 16 h at 37°C, the numbers of red and white colonies were counted.

The suspensions were diluted to 2.5 × 109 CFU/mL, and each diluent (5 μL) was spotted onto Pearlcore MAC plates and cultured for 16 h at 37°C.

Experiment 3. K-12_lacZIS:piee cultured for 8 h in LB at 37°C were inoculated in 2 mL of LB or LB supplemented with 1% (equivalent to 30 mM) lactose, 30 mM IPTG or 0.3 mM IPTG at 0.1 OD600 and grown for 18 h at 37°C with shaking. Clones in the cultures were examined by the same procedures as in Experiments 2-i and 2-ii.

Experiment 4. K-12_lacZIS:piee cultured for 8 h in LB at 37°C were inoculated in 5 mL of LB or LB supplemented with lactose or maltose at 0.1 OD600 and grown for 24 h at 37°C with shaking. At each time point, the OD600 of each culture was determined, and bacterial cells were collected to purify total cellular DNA using the DNeasy blood and tissue kit (Qiagen). Using cellular DNA, the copy number of IS600-excised lacZ was determined by droplet digital PCR using the EvaGreen assay (Bio–Rad) with lacZ-specific primers (5′-AGCCGCTACAGTCAACAGCA-3′ and 5′-ACGCGAAATACGGGCAGACA-3′). The copy number relative to that of lacY was determined by dividing the copy number of IS600-excised lacZ by that of lacY. The lacY gene was amplified with the following specific primers: 5′-AGTAAAACGGCGAGGATGAGTG-3′ and 5′-GCGGATGTTTGGCTGTGTTG-3′.

Experiment 5. K-12_lacZIS:piee cultured for 8 h in LB at 37°C were inoculated into 2 mL of LB supplemented with lactose and several concentrations (0%, 0.1%, 0.2%, and 0.4%) of glucose and grown for 18 h at 37°C with shaking. Clones in the cultures were examined by counting red or white colonies as described for Experiment 2-i.

Experiment 6. K-12_lacZIS:piee was streaked onto MM agar and cultured for 16 h at 37°C. Colonies suspended in MM were inoculated in 5 mL of MM or MM supplemented with lactose or maltose at 0.05 OD600 and grown for 48 h at 37°C with shaking. At each time point, the OD600 of each culture and the copy number of IS600-excised lacZ and lacY in bacterial cells were determined as described for Experiment 4.

Experiment 7. K-12_lacZIS:piee was streaked onto MM agar and cultured for 16 h at 37°C. Colonies suspended in MM were inoculated in 2 mL of MM supplemented with lactose and several concentrations (0%, 0.1%, and 0.2%) of glucose at 0.05 OD600 and grown for 36 h at 37°C with shaking. Clones in the cultures were examined by counting red or white colonies as described for Experiment 2-i.

Analysis of the IS insertion into lacZ and iee among O121:H19 genomes.

Strategies for searching the IS insertion into lacZ and iee are outlined in Fig. S6. Briefly, the lacZ (locus tag: SE14002_0329 in AP024473) and iee (SE14002_1200 in AP024473 and EC51104_3767 in AP024471) genes in 433 O121:H19 draft genomes were first identified by a BLATN-based search. When a gene split by some insertion was detected, we examined the presence and type of inserted IS using the ISCompare program (33) with the ISFinder database (access on Oct. 2020; https://github.com/thanhleviet/ISfinder-sequences) (2).

SNP detection and phylogenetic analysis.

SNP sites (7,591 sites) of the core genome sequences of 419 O121:H19 strains (418 L1 strains and one L3 strain used as the outgroup) were detected by MUMmer (34). After removing recombinogenic SNPs by Gubbins (35), they were used to construct an ML tree in RAxML (36) with the GTR gamma substitution model as previously described (13). ML trees were displayed and annotated using iTOL (37).