Iron Homeostasis Regulates the Genotoxicity of Escherichia coli That Produces Colibactin

Abstract

The genotoxin colibactin is a secondary metabolite produced by a variety of pathogenic enterobacteria. Its biosynthesis requires the enzymatic activity of the phosphopantetheinyl transferase (PPTase) ClbA. We previously showed that ClbA can also contribute to the production of siderophores. Because the biosynthesis of siderophores is regulated by iron availability, we hypothesized that iron could also modulate the production of colibactin through the transcriptional regulation of clbA. This study revealed an increased transcription of clbA under iron-limiting conditions and a decrease of clbA expression in iron-rich media. We demonstrate that clbA transcription is regulated by both the ferric uptake regulator (Fur) and the small regulatory noncoding RNA RyhB. We evidenced that the regulation of the transcription of clbA by Fur and RyhB leads to the regulation of colibactin production. This work highlights the complex mechanism of regulation of an important virulence factor by the two major regulators of bacterial iron homeostasis, making iron a key environmental factor contributing to bacterial virulence and carcinogenesis.

INTRODUCTION

In the human body, most of the iron is intracellular or sequestered by proteins, which makes it unavailable to invading microbes. This process, i.e., nutritional immunity, is a critical host defense strategy against bacterial pathogens that perceive the scarcity of iron within the vertebrate host as a means to sense that they are in their host (1). The global response to iron deprivation in bacteria relies on the ferric uptake regulator (Fur) protein and the small regulatory noncoding RNA (sRNA) RyhB (2–4). In the classical Fur regulation paradigm, Fur binds ferrous ions and the dimeric Fe²⁺-Fur complex (holo-Fur) recognizes target sequences upstream of iron-regulated genes and represses their transcription. However, nowadays numerous reports support four modes of Fur regulation, i.e., apo- and holo-Fur activation and repression, establishing a significant deviation from the classical model of Fur regulation (5). In bacteria, archaea, and eukaryotes, small riboregulators have been shown to mediate posttranscriptional mechanisms of gene regulation (6). In many cases, sRNAs form base pairs with and sequester mRNA ribosome-binding sites, resulting in translational repression and accelerated transcript decay. In contrast, a growing number of examples of translational activation and mRNA stabilization by sRNAs have now been documented. A given sRNA often employs a conserved region to interact with and regulate both repressed and activated targets (6, 7).

To counteract iron deprivation, bacterial pathogens synthesize small iron-scavenging molecules, i.e., siderophores, that are crucial for their survival and play a significant role in virulence (8–10). Siderophores are nonribosomal peptides (NRP) or polyketide (PK)-NRP hybrids (11). A prerequisite for the synthesis of all NRPs, PKs, and PK-NRP hybrids is the posttranslational attachment of P-Pant arms from coenzyme A (CoA) to a conserved serine residue of a carrier protein, converting inactive apo-synthases to active holo-synthases. This reaction is catalyzed by members of the type II family of phosphopantetheinyl transferases (PPTases) (12). The core genome of Escherichia coli codes for two type II PPTases, i.e., EntD and YieE (ECK3705, b3712, JW3690) (13). Whereas the role of YieE remains elusive, EntD is the PPTase involved in the biosynthesis of the siderophores enterobactin, salmochelin, and yersiniabactin.

An additional type II PPTase, i.e., ClbA, is encoded on the pks gene cluster, responsible for the synthesis of the PK-NRP hybrid colibactin in diverse pathogenic enterobacteria, including approximately 50% of the E. coli strains that belong to phylogenetic group B2 (14). Colibactin is a genotoxin that induces DNA double-strand breaks, senescence, and chromosomal abnormalities in eukaryotic cells (15–17) and was demonstrated to be a bona fide virulence factor in a mouse model of sepsis (18) and in a rat model of neonatal meningitis (19). We have recently highlighted the existence of cross talk between colibactin and siderophore biosynthesis (20). Indeed, ClbA can replace EntD and contribute to siderophore biosynthesis. Therefore, we speculated that this interplay could constitute a fitness advantage for E. coli and that iron could play a key role in this connection.

Here we show that the expression of clbA is controlled by iron bioavailability and that the production of colibactin is regulated by Fur and RyhB. Therefore, iron could constitute a key environmental factor contributing to the virulence and carcinogenicity of E. coli strains producing colibactin.

MATERIALS AND METHODS

Bacterial strains, mutagenesis, and growth conditions.

The bacterial strains used in this study are listed in Table 1. For genetic manipulations, E. coli strains were routinely grown at 37°C with shaking at 240 rpm in 3 ml of Lennox L broth (LB; Invitrogen). Ampicillin (50 μg/ml), kanamycin (50 μg/ml), or chloramphenicol (25 μg/ml) was added to the medium when required. Gene inactivations were performed using the bacteriophage lambda Red recombinase method (21) and the primers listed in Table 1. Briefly, the ryhB::cat and fur::kan alleles were PCR amplified from the chromosomal DNA of strain MG1655 ΔryhB or Bw25113 Δfur, respectively, using primers CMD1171_ST/CMD1172_ST or Fur_F/Fur_R, respectively. The purified PCR product was used to transform the strains of interest. The allelic exchanges were confirmed by PCR using the same pairs of primers. The resulting mutants where ryhB or fur was disrupted by a resistance cassette were named the ΔryhB and Δfur mutants, respectively. To generate the fur::cat mutation in strain Nissle 1917 clbA-lux, the chloramphenicol cassette was PCR amplified using primers Fur_P1/Fur_P2. The purified PCR product was used to transform E. coli Nissle 1917 clbA-lux. The transformants were tested by PCR amplification using primers Fur_F/Fur_R. For cloning of the wild-type ryhB and fur genes in cloning vectors, the ryhB and fur genes were PCR amplified using primer pairs RyhB_F_Compl/RyhB_R_Compl and Fur_F_Compl/Fur_R_Compl, respectively. The resulting PCR products were then cloned into the pSC-A vector using a StrataClone PCR cloning kit (Agilent Technologies). The presence of the appropriate insert in the resulting plasmids was checked by PCR amplification, and the plasmids were transformed into the mutant strains.

TABLE 1.

Strains, plasmids, and primers used in this study

Strain, plasmid, or primer	Description or sequencea	Reference or source
E. coli strains
Nissle 1917 clbA-lux	Luciferase fusion of the gene clbA, Kanr	22
Nissle 1917 clbR-lux	Luciferase fusion of the gene clbR, Kanr	22
Nissle 1917 clbB-lux	Luciferase fusion of the gene clbB, Kanr	22
Nissle 1917 clbQ-lux	Luciferase fusion of the gene clbQ, Kanr	22
Nissle 1917 clbA-lux ΔryhB	ryhB mutant strain, Kanr Cmr	This study
Nissle 1917 clbA-lux ΔryhB+p-ryhB	ryhB mutant strain carrying p-ryhB, Kanr Cmr Ampr	This study
Nissle 1917 clbA-lux Δfur	fur mutant strain, Kanr Cmr	This study
Nissle 1917 clbA-lux Δfur+p-fur	fur mutant strain carrying p-fur, Kanr Cmr Ampr	This study
M1/5	Colibactin producer, commensal strain	20
M1/5 ΔryhB	ryhB mutant strain, Cmr	This study
M1/5 ΔryhB+p-ryhB	ryhB mutant strain carrying p-ryhB, Cmr Ampr	This study
M1/5 Δfur	fur mutant strain, Kanr	This study
M1/5 Δfur+p-fur	fur mutant strain carrying p-fur, Kanr Ampr	This study
Nissle 1917	Colibactin producer, probiotic strain	16
Nissle 1917 ΔclbA+pBAD33-clbA		This study
Nissle 1917 ΔclbA+pBAD33-clbA	clbA mutant strain carrying pBAD33-clbA, Kanr	This study
Nissle 1917 ΔclbA+pBAD33-clbAΔclbR	clbA mutant strain carrying pBAD33-clbAΔclbR, Kanr	This study
Nissle 1917 ΔryhB	ryhB mutant strain, Cmr	This study
Nissle 1917 ΔryhB+p-ryhB	ryhB mutant strain carrying p-ryhB, Cmr Ampr	This study
Nissle 1917 Δfur	fur mutant strain, Kanr	This study
Nissle 1917 Δfur+p-fur	fur mutant strain carrying p-fur, Kanr Ampr	This study
SP15	Colibactin producer, pathogenic strain	35
SP15 ΔryhB	ryhB mutant strain, Cmr	This study
SP15 ΔryhB+p-ryhB	ryhB mutant strain carrying p-ryhB, Cmr Ampr	This study
SP15 Δfur	fur mutant strain, Kanr	This study
SP15 Δfur+p-fur	fur mutant strain carrying p-fur, Kanr Ampr	This study
MG1655 ΔryhB	ryhB mutant strain, Cmr	Gift from C. M. Dozois
Bw25113 Δfur	fur mutant strain, Kanr	Gift from C. M. Dozois
Plasmids
p-ryhB	High-copy-no. pSC-A plasmid carrying the ryhB gene; Ampr Kanr	This study
p-fur	High-copy-no. pSC-A plasmid carrying the fur gene; Ampr Kanr	This study
pBAD33-T7-Venus	High-copy-no. pBAD33 plasmid containing the Venus gene coding for a variant of the yellow fluorescent protein of Aequorea victoria	This study
pBAD33-T7-clbA	High-copy-no. pBAD33-T7 plasmid carrying the Venus fusion of the clbA gene	This study
pBAD33-T7-clbAΔclbR	High-copy-no. pBAD33-T7 plasmid carrying the Venus fusion of the clbA gene and mutated clbR	This study
Primers
Fur-P1	AAAATCCTGGAAGTTCTTCAGGAGCCGGACAACCATCACGTCAGTGCGGAGTGTAGGCTGGAGCTGCTTC	This study
Fur-P2	TTCGCGGCAATCGCCTTCGGCACAGTGACCGTAAAGATAGAGACTGTGGTCATATGAATATCCTCCTTAG	This study
Fur_F	CGCCCTAAAGAAAGCTGGCC	This study
Fur_R	CCTTCGTGCGCATGTTCATC	This study
Fur_F_Compl	CTGTAAGCTGTGCCACGTTTT	This study
Fur_R _compl	CTGAGAGCTGTAACTCTCGCTTTTC	This study
CMD1171_ST (ryhB::cat)	TTTGGGGTAAATGTCCCTTTC	29
CMD1172_ST (ryhB::cat)	GTGCGCATAACGAACACAAG	29
RyhB_F_Compl	CCTCTCGAGAAAGCGGACGTGGTTCCTAC	This study
RyhB_R_Compl	GTACTCGAGTGTTTCTGCGTGGCGTATTAC	29
Venus-XbaI	GCGCTCTAGATGGTGTCTATCACTAAAGATCAAATC	This study
Venus-fus-as	GCGCGTCGACTTACTTGTACAGCTCGTCCATG	This study
T7-term-fw	AAGGGGATCCGGCTGCTAAC	This study
T7-term-rev	GCGCAATTGCGGATATAGTTCCTCCTTTCAG	This study
pBAD33-Venus-Bam-rev	GCGCGCGGATCCGGCTGCTAACAAAG	This study
pBAD33-Venus-fw	P-TGGTGTCTATCACTAAAGATCAAATC	This study
clbAp-BamHI-fw	GCGCGGATCCAACCATCACCTTATTATCGG	This study
clbAp-rev1-1	TTAGATAATCTCATTCCTGTTAGC	This study
clbAp-DelclbR-fw2	CCGTTATCTAAGACAAGTATTGCGCATG	This study
clbAp-DelclbR-rev2	ACTTGTCTTAGATAACGGGTTTTTTTCTTTG	This study
fyuA-up	CAACGCGCAGGCCTTTAC	This study
fyuA-rev	GCGTGCTTTCGTCTTGCTG	This study
clbC-up	CGAGTCAAATGCGCCATCAC	This study
clbC-rev	GGACCGCCATACCAATAATG	This study
JPN41	CAGATACACAGATACCATTC	14
JPN42	TCAATGAGGAAGAAATAAAAC	14

^aRestriction sites are underlined. A P in front of the sequence indicates a phosphate residue.

For the megalocytosis assay and assays for bioluminescence and fluorescence measurements, the E. coli strains were grown overnight in Dulbecco's modified Eagle's medium (DMEM) containing HEPES (DMEM-HEPES; Gibco), supplemented with antibiotics when required, at 37°C with shaking (240 rpm). The overnight cultures were then diluted 1:50 in DMEM-HEPES supplemented or not with iron (FeCl₃, 100 μM; FeSO₄, 100 μM), magnesium (MgSO₄, 100 μM), copper (CuCl₂, 100 μM), or iron chelator (8-hydroxyquinoline [8-HQ; 40 μM] or desferrioxamine [DFO; 0.2 μM]) and grown until the optical density at 600 nm (OD₆₀₀) was 0.6. The ranges of concentrations of the chelator and the iron supplement were analyzed, and the highest concentrations of substances which did not alter bacterial growth were chosen.

Construction of Venus fusions.

A low-copy-number vector system was used to generate clbA promoter fusions with the Venus gene as a reporter. For this purpose, a derivative of vector pBAD33, which was devoid of the elements required for l-arabinose-dependent expression, was employed. An overview about the construction of the plasmids is given in Fig. S3 in the supplemental material.

In more detail, the translational fusion of the first 24 bp of rplL and the Venus gene, coding for a variant of the yellow fluorescent protein of Aequorea victoria, was amplified from pMBrplL-Venus using the primers Venus-XbaI and Venus-fus-as. The PCR product was digested with XbaI and SalI and cloned into pBAD33 (see step 1 in Fig. S3 in the supplemental material). The resulting plasmid was digested with EcoRV and SacI, treated with mung bean nuclease (MBN), and religated as plasmid pBAD33-Venus in order to delete the araBAD promoter and the araC gene (see step 2 in Fig. S3). Since the aim was to clone the whole clbBR intergenic region together with the clbR gene, which harbors the clbA promoter, the transcription into the opposite direction of clbA, facing clbB, had to be blocked. To achieve this, the T7 terminator sequence of pTXB1 (NEB) was amplified with primers T7-term-for and T7-term-rev, subsequently digested with BamHI and MfeI, and cloned upstream of Venus into pBAD33-Venus, resulting in pBAD33-T7-Venus (see step 3 in Fig. S3). The plasmid was verified by sequencing.

The whole pBAD33-T7-Venus sequence was amplified via PCR as a precursor for the following plasmids by using primers pBAD33-Venus-Bam-rev and pBAD33-Venus-fw and subsequently digested with BamHI (see step 4 in Fig. S3). To yield pBAD33-T7-clbA, the 648-bp DNA sequence upstream of clbA in E. coli strain M1/5 (20), comprising the putative clbA promoter, the clbR gene, as well as the whole clbBR intergenic region (clbBR in. reg.), was amplified with primers clbAp-BamHI-fw and clbAp-rev1-1, digested with BamHI, and then ligated with the digested pBAD33-T7-Venus amplicon (see step 5a in Fig. S3). To generate pBAD33-T7-clbAΔclbR, a derivative of pBAD33-T7-clbA with a deletion of clbR nucleotides 33 to 40, a two-step-PCR was carried out using E. coli M1/5 as a template (see step 5b in Fig. S3). Amplicons were generated with the primer pairs clbAp-BamHI-fw/clbAp-DelclbR-rev2 (PCR 1) and clbAp-DelclbR-fw2/clbAp-rev1-1 (PCR 2). PCR products 1 and 2 were fused using primers clbAp-BamHI-fw and clbAp-rev1-1 (PCR 3) and then digested with BamHI and ligated with PCR-amplified and BamHI-digested pBAD33-T7-Venus, resulting in pBAD33-T7-clbAΔclbR. The plasmids were verified by sequencing.

Luciferase and Venus measurements.

The promoter activities of genes clbA, clbB, clbQ, and clbR in different media were determined by time course quantification of luciferase or Venus expression. When an OD₆₀₀ of 0.6 was reached, bacterial subcultures were diluted to an OD₆₀₀ of 0.1 in the appropriate medium. Samples of 100 μl were then used to inoculate a black 96-well plate (Greiner Bio-One), and the bacteria were grown without shaking at 37°C in a luminometer (Tecan Infinite Pro microplate reader). For luminescence measurements, light emission (relative light units [RLU]) was recorded (6,000-ms aperture per sample) every 30 min in parallel with the OD₆₀₀. For Venus fluorescence measurements, fluorescence was recorded at 535 nm (bandwidth, 25 nm) using an excitation wavelength of 485 nm (bandwidth, 20 nm) with the luminometer. The OD₆₀₀ was measured in parallel at each time point. The area under the curve (AUC), which quantifies the cumulative luminescence or fluorescence, was calculated with GraphPad Prism (version 6.0) software.

Quantification of colibactin-associated genotoxic effect by megalocytosis assay.

Quantification of the colibactin-associated genotoxic effect by megalocytosis assay was performed as previously described (14). Briefly, HeLa cells were grown in DMEM (GlutaMAX; Invitrogen) supplemented with 10% (vol/vol) fetal calf serum (FCS; Eurobio) and 1% (vol/vol) nonessential amino acids (Invitrogen) at 37°C in a 5% CO₂ atmosphere. HeLa cells were dispensed in a 96-well cell culture plate (5 × 10³ cells/well) for 24 h and then infected at a multiplicity of infection (MOI; number of bacteria per HeLa cell at the onset of infection) of 100 with the E. coli strains, which had previously been pregrown in the appropriate medium. At 4 h postinoculation, the cells were washed 3 times with Hanks' balanced salt solution (HBSS; Gibco) and incubated in cell culture medium for 72 h with 200 mg/ml gentamicin before fixation (4% formaldehyde) and protein staining with methylene blue (1% [wt/vol] in 0.01 M Tris-HCl). The methylene blue was extracted with 0.1 N HCl. Staining was quantified by measurement of the OD₆₆₀.

Electrophoretic mobility shift assay (EMSA).

PCR products were generated by using genomic DNA from E. coli strain M1/5 as the template and the primers JPN42/JPN41 (clbA, 343 bp), clbC-up/clbC-rev (clbC, 219 bp), and fyuA-up/fyuA-rev (fuyA, 648 bp) (Table 1). The PCR products were gel purified using a Wizard SV gel and PCR cleanup system (Promega). A 20-μl binding reaction mixture containing recombinant Fur protein (MyBioSource, USA), 50 μg ml⁻¹ of poly(dI-dC), 4 μl of 5× binding buffer (50 mM Tris-HCl, pH 7.5, 200 mM KCl, 5 mM MgCl₂, 0.625 mM MnCl₂, 25% [vol/vol] glycerol, 2.5 mM dithiothreitol, 500 μg ml⁻¹ bovine serum albumin), and the DNA probe (10 nM) was incubated at room temperature for 30 min, loaded onto a 4% Tris-borate polyacrylamide gel, and electrophoresed in 0.5× Tris-borate, pH 7.5, containing 0.2 mM MnCl₂ at 100 V. The gels were stained with ethidium bromide.

Statistical analysis.

Statistical analyses were conducted using GraphPad Prism (version 6.0) software. The mean with the standard error of the mean (SEM) is shown in the figures, and P values were calculated using a one-way or two-way analysis of variance (ANOVA) followed by a Bonferroni posttest, unless otherwise stated. P values of less than 0.05 were considered statistically significant.

RESULTS

Iron availability modulates clbA transcription.

In order to test whether the promoter activities of the genes located in the pks island (see Fig. S1A in the supplemental material) were dependent on iron availability, E. coli strain Nissle 1917 harboring transcriptional luciferase fusions with the genes clbA, clbB, clbQ, and clbR, four genes involved in colibactin biosynthesis, were studied (Table 1; see also Fig. S1A and B) (22). These strains were grown in chemically defined DMEM-HEPES, DMEM-HEPES supplemented with 100 μM FeCl₃, and DMEM-HEPES supplemented with an iron chelator, 8-hydroxyquinoline (8-HQ; 40 μM). The ranges of concentrations of the chelator and the iron supplement were analyzed, and the highest concentrations of substances which did not alter bacterial growth were chosen. The transcription rates of the genes clbA, clbB, clbQ, and clbR were determined as the number of OD₆₀₀-standardized relative luminescence units (RLU) (Fig. 1).

FIG 1.

Iron availability modulates the transcription of clbA. Growth curves (OD₆₀₀) and the RLU/OD₆₀₀ of Nissle 1917 clbA-lux (A), clbB-lux (B), clbQ-lux (C), and clbR-lux (D) fusion strains grown at 37°C in DMEM-HEPES, DMEM-HEPES supplemented with 40 μM 8-HQ, or DMEM-HEPES supplemented with 100 μM FeCl₃. (E) AUC, determined with GraphPad Prism (version 6.0) software, of the RLU/OD₆₀₀ of clbA under the three different conditions. The given values are the mean number of RLU and SEMs from five independent experiments. Statistical analysis was performed using two-way ANOVA and the Bonferroni posttest. ***, P < 0.001 compared to the mean values obtained in DMEM-HEPES.

This analysis revealed that only clbA was significantly transcribed in DMEM-HEPES, as previously reported (22) (compare Fig. 1A to Fig. 1B to D). During the bacterial growth kinetics, a continuous increase in bioluminescence emission was observed to reach a maximal value at early stationary phase of growth, followed by a decrease of transcription (Fig. 1A).

The transcription of the genes clbB, clbQ, and clbR was not significantly modified when the culture medium was iron depleted (DMEM-HEPES supplemented with 40 μM 8-HQ) or iron replete (DMEM-HEPES supplemented with 100 μM FeCl₃) (Fig. 1B, C, and D, respectively). In contrast, under iron-depleted conditions, the expression of clbA was enhanced and maintained longer than that in DMEM-HEPES (Fig. 1A). Under iron-supplemented conditions, clbA transcription was strongly repressed (Fig. 1A). Calculation of the area under the curve (AUC) of the relative number of OD₆₀₀-standardized relative luminescence units (RLU/OD₆₀₀) confirmed the iron-dependent differential clbA expression (Fig. 1E). Supplementation with an alternative iron chelator (desferrioxamine [DFO], 0.2 μM) or an alternative iron source (FeSO₄, 100 μM) resulted in similar patterns (see Fig. S2 in the supplemental material).

In order to test whether other divalent metals could influence clbA transcription, we monitored the expression of clbA in strain Nissle 1917 clbA-lux grown in DMEM-HEPES supplemented with copper (CuCl₂, 100 μM) or magnesium (MgSO₄, 100 μM) (Fig. 2). This revealed that the transcription of clbA was not altered in the presence of Mg²⁺ or Cu²⁺.

FIG 2.

The transcription of clbA is not modulated by Cu²⁺ or Mg²⁺. Growth curves (OD₆₀₀) and the RLU/OD₆₀₀ of E. coli Nissle 1917 clbA-lux grown at 37°C in DMEM-HEPES and DMEM-HEPES supplemented with either 100 μM MgSO₄ or 100 μM CuCl₂ (A) and the associated AUC (B). The given values are the mean number of RLU and SEMs from at least three independent experiments. Statistical analysis was performed by comparison of the values to the mean values obtained in DMEM-HEPES using two-way ANOVA and the Bonferroni posttest. ns, not significant.

Altogether, these results show that clbA transcription is regulated by iron availability in the medium.

The iron-dependent transcription of clbA is independent of the LuxR-like protein ClbR.

The gene clbR encodes a LuxR-like protein that exhibits a helix-turn-helix DNA-binding motif, which is suspected to be a regulator of colibactin biosynthesis (22). Moreover, cotranscription was previously demonstrated for the genes clbR and clbA in DMEM-HEPES medium with 5% fetal bovine serum (22). Our results described above showed that only clbA transcription and not clbR transcription is regulated by iron. To test whether ClbR is involved in the iron-dependent regulation of clbA, the clbA gene was transcriptionally fused with an alternative reporter gene, i.e., the Venus gene (coding for a variant of yellow fluorescent protein of Aequorea victoria) in plasmid pBAD33-T7-Venus to produce pBAD33-T7-clbA. The clbR gene was inactivated to produce plasmid pBAD33-T7-clbAΔclbR. The resulting plasmids, pBAD33-T7-clbA and pBAD33-T7-clbAΔclbR, were transformed into strain Nissle 1917 ΔclbA. Monitoring of the fluorescence of the Venus protein transcriptionally fused with the clbA gene in strain Nissle 1917 ΔclbA+pBAD33-T7-clbA confirmed the iron-dependent regulation of clbA (Fig. 3A and B). Monitoring of the fluorescence of strain Nissle 1917 ΔclbA+pBAD33-T7-clbAΔclbR under iron-supplemented and iron-chelated conditions revealed that clbA transcription was regulated by the iron concentration in the absence of functional ClbR (Fig. 3A and C). These results show that ClbR is not involved in the iron-dependent regulation of clbA.

FIG 3.

The transcription of clbA is dependent on the iron concentration in a ClbR-independent manner. (A) Sequence of the construction of the translational Venus fusion of clbA (clbA-Venus fusion). Two different constructs, one containing the wild-type clbR gene and one containing an inactive clbR gene, were generated. The underlined sequence was mutated to inactivate clbR. The wild-type and the mutated constructs were cloned into plasmid pBAD33-T7, and the plasmid was transformed into strain Nissle 1917 in which clbA was mutated. (B and C) Schematic representation of the plasmid construct obtained and results of the quantification of fluorescence of strain Nissle 1917 in which clbA was mutated and transformed with the plasmid containing the wild-type construct (B) or the mutated one (C). The fluorescence was measured in DMEM-HEPES or DMEM-HEPES supplemented with 100 μM FeCl₃ or 40 μM 8-HQ. The given values are the mean fluorescence/OD₆₀₀ and SEMs three independent experiments. Statistical analysis was performed by comparison of the values to the mean values obtained in DMEM-HEPES using two-way ANOVA and the Bonferroni posttest. ***, P < 0.001; **, P < 0.01.

Fur positively regulates the production of colibactin.

Because the global response to iron availability in bacteria relies on the ferric uptake regulator (Fur) protein, we tested whether the iron-dependent regulation of clbA transcription was mediated by Fur. We mutated the fur gene in strain Nissle 1917 clbA-lux. The complemented strain Nissle 1917 clbA-lux Δfur+p-fur, where the plasmid-borne wild-type fur gene was transformed into strain Nissle 1917 clbA-lux Δfur, was also constructed (Table 1). Monitoring of the expression of clbA in the resulting strains (Fig. 4A and B) revealed that the transcription of clbA was significantly decreased in the strain in which fur was mutated (Fig. 4A and B). Transformation of the strain in which fur was mutated with a plasmid carrying the functional wild-type fur gene (Fig. 4A and B) totally restored the expression of clbA. This indicates that the transcription of clbA is positively regulated by Fur.

FIG 4.

Fur regulates the transcription of clbA and colibactin biosynthesis. (A) Growth curves (OD₆₀₀) and the RLU/OD₆₀₀ of strains Nissle 1917 clbA-lux, Nissle 1917 clbA-lux Δfur, and Nissle 1917 clbA-lux Δfur+p-fur grown at 37°C in DMEM-HEPES. The given values of the number of RLU and SEMs resulted from three independent experiments. (B) AUC of the RLU/OD₆₀₀ of clbA in the three different strains. (C) Quantification of colibactin production. The production of colibactin by the E. coli strains and derivatives was determined by quantification of megalocytosis as previously described (15). The multiplicity of infection was 100 bacteria per cell. Three distinct genetic contexts of E. coli were investigated: Nissle 1917 (a probiotic strain), M1/5 (a commensal strain), and SP15 (a strain isolated from a patient with neonatal meningitis). Both cells that were not infected (NI) and cells that were infected with E. coli strains that do not produce colibactin gave the same quantitative results (data not shown). The given quantification values are represented as the mean values and SEMs from three independent experiments. Statistical analysis was performed by comparison of the values to the mean values obtained for wild-type (WT) strains using two-way ANOVA and the Bonferroni posttest. ***, P < 0.001; **, P < 0.01; *, P < 0.05; ns, not significant.

Since ClbA is required for the production of colibactin, the role of Fur in the synthesis of colibactin was then investigated. Eukaryotic cells were infected with pks⁺ E. coli strains in which fur was mutated or not. Three different genetic contexts were analyzed: strains Nissle 1917 (a probiotic strain), M1/5 (a commensal strain), and SP15 (a pathogenic strain isolated from a patient with neonatal meningitis) (Table 1). The genotoxic effect of colibactin was monitored by quantification of megalocytosis, as previously described (14). This revealed that all the Δfur mutants displayed a genotoxicity significantly reduced compared to that of the wild-type strains (Fig. 4C). Transformation of the strains in which fur was mutated with a plasmid carrying the functional wild-type fur gene resulted in restoration of the megalocytosis phenotype (Fig. 4C). Altogether, these results evidence that Fur positively regulates colibactin production.

To determine whether Fur could regulate clbA directly and/or indirectly, we investigated a putative direct binding of Fur to clbA. We detected two putative Fur-binding sites in the DNA sequence of the clbA promoter (Fig. 5A). We tested whether purified Fur protein bound to its putative binding sites in the clbA promoter region (Fig. 5B). EMSA showed that migration of the 343-bp DNA fragment was retarded when it was incubated with the Fur protein. This suggests that Fur can regulate clbA expression through direct binding.

FIG 5.

Fur binds to the clbA promoter. (A) Identification and localization of the two putative Fur boxes located upstream of clbA. (B) EMSA was performed in the presence of 0 μM (lanes a), 0.25 μM (lanes b), 0.5 μM (lanes c), 1 μM (lanes d), and 4 μM (lanes e) Fur protein and 10 ng of the PCR product amplified with primers JPN41/JPN42. Three PCR products were tested: a negative control (clbC, 229 bp), a positive control (fyuA, 648 bp), and the promoter region upstream of clbA (343 bp).

RyhB negatively regulates the production of colibactin.

Because the second major regulator of bacterial iron metabolism is the small regulatory noncoding RNA RyhB, we investigated whether RyhB was involved in the iron-mediated regulation of clbA transcription. The gene ryhB was inactivated in strain Nissle 1917 clbA-lux, and the complemented strain was also engineered (Table 1). Monitoring of the bioluminescence in the resulting strain, Nissle 1917 clbA-lux ΔryhB, revealed that the level of transcription of clbA was enhanced and maintained longer than it was in strain Nissle 1917 clbA-lux (Fig. 6A and B). Overexpression of the wild-type ryhB gene expressed from a high-copy-number plasmid in strain Nissle 1917 clbA-lux ΔryhB resulted in the repression of clbA transcription (Nissle 1917 clbA-lux ΔryhB+p-ryhB) (Fig. 6A and B). These data evidence that RyhB plays a critical role in the transcription of clbA.

FIG 6.

RyhB regulates the transcription of clbA and colibactin biosynthesis. (A) Growth curves (OD₆₀₀) and the RLU/OD₆₀₀ of strains Nissle 1917 clbA-lux, Nissle 1917 clbA-lux ΔryhB, and Nissle 1917 clbA-lux ΔryhB+p-ryhB grown at 37°C in DMEM-HEPES. The given values are the mean number of RLU and SEMs from three independent experiments. (B) AUC of the RLU/OD₆₀₀ of clbA in the three different strains. (C) Quantification of colibactin production. The production of colibactin by E. coli strains and its derivatives was determined by quantification of megalocytosis as previously described (14). The multiplicity of infection was 100 bacteria per cell. Three distinct genetic contexts of E. coli were investigated: Nissle 1917 (a probiotic strain), M1/5 (a commensal strain), and SP15 (a strain isolated from a patient with neonatal meningitis). Both cells that were not infected (NI) and cells that were infected with E. coli strains that do not produce colibactin gave the same quantitative results (data not shown). The given quantification values are represented as mean values and SEMs from three independent experiments. Statistical analysis was performed by comparison of the values to the mean values obtained for wild-type strains using two-way ANOVA and the Bonferroni posttest. ***, P < 0.001; **, P < 0.01; *, P < 0.05.

The role of RyhB in the synthesis of colibactin was then investigated. Eukaryotic cells were infected with pks⁺ E. coli strains in which ryhB was mutated or not. The genotoxic effect of colibactin, quantified by megalocytosis, revealed that all the ΔryhB mutants displayed significantly increased genotoxicity compared to the wild-type strains (Fig. 6C). Overexpression of ryhB in strains in which the ryhB gene was mutated resulted in a decrease in the level of the colibactin-associated genotoxic effect (Fig. 6C). Altogether, these results evidence that RyhB negatively regulates colibactin production under these conditions.

To determine whether RyhB could regulate clbA directly and/or indirectly, we investigated a putative direct pairing of RyhB to the clbA promoter. The RNA Hybrid computational program (23) predicted a region of potential pairing between RyhB and clbA (Fig. 7). The region of pairing with clbA was located 238 to 344 nucleotides downstream from the translational start (minimal free energy, −51.3 kcal/mol) (Fig. 7) in the reading frame of clbA. This suggests that ryhB could regulate clbA through direct binding.

FIG 7.

RyhB seems to interact directly with the clbA gene. The RNA Hybrid computational program predicted a region of potential interaction between RyhB and clbA. The nucleotide sequence involved in the interaction between clbA (gray) and RyhB (black) (A) and the two-dimensional structure of RyhB (green) bound to clbA (red) (B) are shown. The predicted pairing region with RyhB began at nucleotide 238 in the clbA open reading frame (minimal free energy, −51.3 kcal/mol).

DISCUSSION

This work highlights that the transcription of clbA, the gene encoding the PPTase involved in colibactin production, is regulated by iron availability (see our model in Fig. 8). Therefore, this work provides new insights into the transcriptional regulation of clbA, as only one study on its regulation has been published previously (22). In addition to carbon source and growth phase (22), our work demonstrates that clbA is regulated by iron availability. Moreover, this regulation appears to be independent of ClbR, the LuxR-like protein predicted to regulate colibactin biosynthesis (22), although the clbR gene was demonstrated to be coexpressed with clbA under particular culture conditions (22). This suggests that clbA could be expressed from a second promoter specifically activated under specific conditions of iron availability. This iron-dependent regulation of clbA transcription could constitute a fitness advantage for the bacteria, as we previously demonstrated that ClbA can also sustain siderophore biosynthesis (20).

FIG 8.

Model for the regulation of the transcription of clbA by Fur and RyhB. The biosynthesis of the genotoxin colibactin requires the enzymatic activity of the PPTase ClbA. The transcription of clbA is dependent on both the ferric uptake regulator (Fur) and the small regulatory noncoding RNA RyhB. This makes iron a key environmental factor contributing to the regulation of colibactin production.

This work highlights that Fur is positively involved in the production of the virulence factor colibactin through the regulation of clbA. Fur is a global transcriptional regulator that controls the transcription of over 90 genes involved in iron uptake, storage, and metabolism. Only a few studies have reported the involvement of Fur in the regulation of the biosynthesis of toxins in E. coli. The stx genes, encoding Shiga toxin in enterohemorrhagic E. coli (24, 25), and hly plasmids, encoding hemolysin (26), are directly regulated by Fur. In addition to its well-known role as a repressor, Fur was characterized as a positive regulator of gene expression, such as the expression of acnA, fumA, ftnA, bfr, and sodB (27).

Our work also evidenced that RyhB regulates clbA transcription, as we observed that clbA transcription was upregulated when the ryhB gene was inactivated, demonstrating that clbA is downregulated by RyhB. RyhB is a small RNA that recruits RNase E and facilitates the degradation of mRNA targets (27, 28). Whereas only a few studies have demonstrated the impact of RyhB on virulence in pathogenic E. coli, for instance, through siderophore production (29), several studies have demonstrated that this sRNA is implicated in the virulence-associated processes of other pathogenic bacteria, such as Shigella flexneri (2) or Shigella dysenteriae (30). This study provides new evidence that RyhB is involved in the production of a virulence factor in pathogenic E. coli through the regulation of clbA transcription, which leads to the modulation of colibactin production.

Here we demonstrated that modulation of the expression of clbA via regulators of iron homeostasis leads to the modulation of colibactin production. The mutation of ryhB in a colibactin-producing E. coli strain leads to an increase in the level of colibactin production, whereas the mutation of fur results in a decrease in the level of colibactin production in three different genetic contexts of E. coli, in which the strains belonged to the B2 phylogenetic group. We confirmed that this regulation occurred in three different strains; however, the quantitative differences observed could be explained by the genetic diversity between strains. Colibactin was demonstrated to be a virulence factor, for example, in neonatal systemic infections (19), but it is also associated with the development of colorectal cancer (31–33). The fact that colibactin is regulated by the two major regulators of iron homeostasis could link iron availability in the gut and blood and E. coli-mediated carcinogenesis and systemic infections, respectively. Moreover, the main reservoir of E. coli is the intestine and, more precisely, the colon. A recent publication confirmed the emergence of the B2 group of E. coli in developed countries (34). Given the existence of a gradient of iron concentration from the lumen to the intestinal epithelial cell, it is conceivable that the fine-tuning of clbA expression allows the production of colibactin when the pathogenic E. coli isolate is located in an appropriate site in the gut. This iron-dependent regulation could be one reason for the emergence of the B2 group E. coli strains, as we know that there is cross talk between colibactin and siderophores. Integration of the regulation of virulence factors, such as siderophores and colibactin, into networks that respond to specific environmental signals, such as the local iron concentration and the balance between Fur and RyhB, could result in the accurate production of colibactin and siderophores, so that the bacteria can adapt to the competitive environment in the gut.

Further understanding of how E. coli senses environments to coordinate the expression of clbA and the mechanism of regulation of the biosynthesis of these virulence factors may uncover novel pathways for the development of potential targets against this pathogenic enterobacterium.