ABSTRACT
In this study, we isolated a bacteriophage T7-resistant mutant strain of Escherichia coli (named S3) and then proceeded to characterize it. The mutant bacterial colonies appeared to be mucoid. Microarray analysis revealed that genes related to colanic acid production were upregulated in the mutant. Increases in colanic acid production by the mutant bacteria were observed when l-fucose was measured biochemically, and protective capsule formation was observed under an electron microscope. We found a point mutation in the lon gene promoter in S3, the mutant bacterium. Overproduction of colanic acid was observed in some phage-resistant mutant bacteria after infection with other bacteriophages, T4 and lambda. Colanic acid overproduction was also observed in clinical isolates of E. coli upon phage infection. The overproduction of colanic acid resulted in the inhibition of bacteriophage adsorption to the host. Biofilm formation initially decreased shortly after infection but eventually increased after 48 h of incubation due to the emergence of the mutant bacteria. Bacteriophage PBECO4 was shown to infect the colanic acid-overproducing mutant strains of E. coli. We confirmed that the gene product of open reading frame 547 (ORF547) of PBECO4 harbored colanic acid-degrading enzymatic (CAE) activity. Treatment of the T7-resistant bacteria with both T7 and PBECO4 or its purified enzyme (CAE) led to successful T7 infection. Biofilm formation decreased with the mixed infection, too. This procedure, using a phage cocktail different from those exploiting solely receptor differences, represents a novel strategy for overcoming phage resistance in mutant bacteria.
INTRODUCTION
With the appearance of antibiotic-resistant pathogenic bacteria, bacteriophage therapy is considered a likely alternative to antibiotics (1–4). The advantages of phage therapy include autodosing (7), minimal influence on normal flora, the ability to control biofilms, and low inherent toxicity (5–8). Although many successful cases of control of bacterial infections with phages have been reported (9), one problem associated with phage therapy is the emergence of phage-resistant mutant bacteria. Two representative mechanisms of bacteria for resisting infection by phage are the prevention of phage adsorption and the cutting of the invading phage genome (10). The former may involve the masking of the phage receptor, the production of a competitive inhibitor, or the production of an extracellular matrix (7, 11). The latter may involve restriction modification or the CRISPR (clustered regularly interspaced short palindromic repeat)-Cas (CRISPR-associated) system (10, 12, 13). Accordingly, current phage therapy generally adopts a cocktail strategy in which phages targeting different cellular receptors are mixed (6).
One of the extracellular matrices produced by bacteria for resisting phage adsorption is colanic acid. Colanic acid is an exopolysaccharide (EPS) that protects bacteria of the family Enterobacteriaceae and acts by forming a protective capsule. It is produced upon osmotic shock, dehydration, or destabilization of the outer membrane (14, 15). In Escherichia coli, genes related to colanic acid production are clustered in the wca (cps) region (16, 17). The Rcs phosphorelay system, comprising the RcsB-RcsA heterodimer, induces the transcriptional activation of these genes (18). Lon protease is a negative regulator degrading the dimer, and RcsA itself is a transcriptional activator with an autoregulation mechanism (19, 20).
In this study, we found a phage T7-resistant mutant of E. coli that overproduces colanic acid. We have analyzed the mechanism for the resistance of this mutant. Here we present a novel method for the successful infection of resistant bacteria with the aid of a second phage harboring a colanic acid-degrading enzyme.
MATERIALS AND METHODS
Bacterial strains and growth conditions.E. coli K-12 (ATCC 23724) and E. coli O157:H7 (ATCC 700927) were used as the host strains for the bacteriophages. Clinical isolates of E. coli were obtained from patients in the Samsung Medical Center, Sungkyukwan University School of Medicine, Suwon, South Korea. Bacteria were grown in Luria-Bertani (LB) medium at 37°C.
Bacteriophage isolation and purification.Phages T7, T4, and lambda were provided by Denise Tremblay of the University of Laval, Quebec, Canada. The isolation and characterization of phage PBECO4 have been described previously (21). The phages were purified using centrifugation through a glycerol gradient method (22). A single plaque was isolated and was used to inoculate 5 ml of a mid-exponential-phase culture of the bacterium, followed by incubation at 37°C for 3 h. A phage lysate was obtained by centrifugation at 11,000 × g for 10 min. Five milliliters of the lysate was used to inoculate 100 ml of a mid-exponential-phase culture of the bacterium, and the mixture was then incubated until lysis was completed. NaCl was added to the lysate at a final concentration of 1 M, and the lysate was incubated at 4°C for 1 h. After centrifugation at 11,000 × g for 10 min, 10% (wt/vol) polyethylene glycol 8000 (PEG 8000) was added, and the mixture was then incubated at 4°C for 1 h. The supernatant was discarded after centrifugation at 11,000 × g for 10 min. Then the pellet was resuspended in 750 μl of SM buffer (100 mM NaCl, 8 mM MgSO4·7H2O, 50 mM Tris-Cl [pH 7.5]), and chloroform was added at a ratio of 1:1 (vol/vol), followed by vortexing and then centrifugation at 3,000 × g for 15 min. The upper phase was isolated and was added to a polycarbonate centrifuge tube containing 3 ml of 40% glycerol in the lower layer and 4 ml of 5% glycerol in the upper layer. After centrifugation at 151,000 × g for 1 h, the supernatant was discarded, and the pellet was resuspended in 400 μl of SM buffer. Purified phages were stored at 4°C until use.
Isolation of phage-resistant mutants.Two hundred microliters of E. coli K-12 grown to mid-exponential phase was mixed with bacteriophage T7 at a multiplicity of infection (MOI) of 10. Four milliliters of top agar (1% tryptone, 0.5% yeast extract, 0.5% NaCl, and 0.7% agar) was added to the mixture, which was poured onto an LB agar plate (1% tryptone, 0.5% yeast extract, 0.5% NaCl, and 1.5% agar). After 24 h of incubation at 37°C, lysis was observed, and some resistant mutant colonies appeared on the plate. Mutant colonies were picked and were transferred to a fresh LB agar plate. They were restreaked to check the stability of the phenotype.
Microarray analysis.Total bacterial RNA was isolated from both wild-type E. coli K-12 and the phage-resistant E. coli mutant S3 using a NucleoSpin RNA II kit (Macherey-Nagel, USA). Transcription profile analysis was carried out using an Affymetrix GeneChip E. coli Genome 2.0 array (DNA Link, South Korea).
Real-time PCR.Reverse transcription was performed using random hexamers on RNA templates isolated from wild-type or phage-resistant mutant E. coli. PCR was performed using Sybr green and a MyiQ single-color real-time PCR detection system (Bio-Rad, USA). The primers used were rcsA F (5′-CGCCCTTCAGTGGTGTTTAT-3′) and rcsA R (5′-GCCAAGGATATCGTCGAGAG-3′), lon F (5′-CTCTCTGACAATGGCGAACA-3′) and lon R (5′-CAGCACTTCTGGTGGGATTT-3′), wza F (5′-TGTCAGCATAGCTGCATTCC-3′) and wza R (5′-CGTTTCTCCAGTCAGCATCA-3′), and GAPDH F (5′-AGGTCTGATGACCACCGTTC-3′) and GAPDH R (5′-GGAACGCCATACCAGTCAGT-3′).
Phage adsorption test.Ten milliliters of bacterial culture was grown to the mid-log phase (optical density at 600 nm [OD600], 0.5), and phages were added at an MOI of 0.1 (5 × 108 PFU). The mixture was incubated at 37°C, and 0.1 ml of culture was obtained at 0, 3, 6, 9, and 12 min postinfection. Free phage was titrated using a standard plaque assay (23).
Electroporation of phage genomic DNA.Bacteria were grown to the mid-exponential phase and were incubated on ice for 30 min. After centrifugation (6,500 × g) at 4°C for 15 min, the supernatant was discarded and the pellet was resuspended in ice-cold distilled water of the same volume as the culture. After another 15 min of incubation on ice, the mixture was subjected to centrifugation (6,500 × g) at 4°C for 15 min. The supernatant was discarded, and the pellet was resuspended in a 1/20 culture volume of ice-cold 10% glycerol, followed by incubation on ice for 15 min. The mixture was subjected to centrifugation (2,500 × g) at 4°C for 10 min, and the supernatant was discarded. The pellet was resuspended in 1/100 volume of ice-cold 10% glycerol to make competent cells. Phage genomic DNA was isolated from 1 × 109 PFU of phages using a Phage DNA isolation kit (Norgen Biotek Corp., Thorold, Ontario, Canada). Two micrograms of phage DNA was mixed with 100 μl of competent cells, and the mixture was poured onto a prechilled 0.2-cm cuvette (Bio-Rad, USA). Electroporation was conducted at 2.5 kV, 25 μF, and 200 Ω with a 5.0-ms pulse.
Purification of CAE.Open reading frame 547 (ORF547) of phage PBECO4 (nucleotides 339997 to 342138) was PCR amplified with forward (F) (5′-CCGAGGATCCATGGCAGTATTTTC-3′) and reverse (R) (5′-GCACGAGCTCTTATGGTAGGGTTATG-3′) primers. The PCR product was cloned into the pQE30 vector between the BamHI and PstI sites (Qiagen). The resulting plasmid was used to transform E. coli BL21(DE3) (Stratagene). After overexpression of the gene, the colanic acid-degrading enzyme (CAE) was purified using Ni-nitrilotriacetic acid (NTA) affinity chromatography (Qiagen). Elution was carried out with 1 ml of elution buffer with 250 mM imidazole.
Reporter assay.To measure the promoter activity of lon, the firefly luciferase gene was cloned into pACYC184 between the BamHI and HindIII sites. Either a wild-type or a mutant lon promoter was amplified with F (5′-AGTTGGATCCAATTCCGTGCACGACGAG-3′) and R (5′-GTACGGATCCAGGTAATCAGATGAC-3′) primers. The PCR product was cloned at the BamHI site. E. coli DH5α was transformed with each reporter plasmid and was grown in 10 ml of LB medium. The bacterial pellets obtained after centrifugation were resuspended in 100 μl of lysis buffer (1 mg/ml lysozyme, 20% [wt/vol] sucrose, 30 mM Tris-Cl [pH 8.0], and 1 mM EDTA). The mixture was incubated on ice for 10 min and was subjected to centrifugation. The supernatant was obtained, and luciferase activity was measured using a Luciferase Reporter Assay system (Promega, USA) and a Fusion luminometer (Packard, USA).
Quantification of colanic acid.In order to quantify colanic acid, a specific component, l-fucose, was measured (24). The bacterial culture was first boiled for 20 min and then cooled down to room temperature. The culture was subjected to centrifugation at 16,000 × g for 20 min. The supernatant was recovered and was diluted 1/10 (vol/vol) with distilled water. A 4.5-ml volume of H2SO4·H2O (6:1, vol/vol) was added to 1 ml of the diluted supernatant, and the mixture was boiled for 20 min. Absorbance at the corresponding wavelength (A-co) was measured at 396 and 427 nm. One hundred microliters of 1 M cysteine hydrochloride was added to the mixture, and absorbance after the addition of 1 M cysteine hydrochloride (A-cy) was measured at 396 and 427 nm. The concentration of l-fucose was calculated as (A396-cy − A396-co) − (A427-cy − A427-co). A standard curve was obtained by measuring l-fucose in the range of 0 to 100 mg/ml using the same method.
Biofilm assay.The method for measuring a biofilm on a plastic surface has been described previously (25, 26). Briefly, bacteria were inoculated into 1 ml LB broth contained in a well of a polystyrene 96-well plate at a concentration of 1 × 106 CFU/ml. After incubation for 24 h, the culture was discarded so as to remove planktonic cells, and the well was washed twice with phosphate-buffered saline. One milliliter of methanol was added to the well for fixation, and incubation was carried out for 15 min. The methanol was discarded, and staining was performed with a 0.4% crystal violet solution for 15 min. For destaining, the well was washed with tap water, and 33% acetic acid was added, with further incubation for 20 min. Absorbance was measured at 595 nm.
Statistical analysis.Student's t test was used to calculate the significance of the differences where indicated. Statistical significance was defined as a P value of <0.05.
RESULTS
Isolation of T7-resistant mutant bacteria.After the infection of the E. coli K-12 bacteria on the agar plates with T7, many phage-resistant mutant colonies arose, some of which appeared to be mucoid. We selected one of these mucoid colonies and named it S3. Transcription profile analysis comparing wild-type and S3 mutant bacteria revealed that genes related to colanic acid production were generally upregulated in the mutant (Table 1). It is noteworthy that the expression of rcsA, a positive regulator of colanic acid production-related gene clusters, increased, while the expression of lon, a negative regulator of rcsA, decreased. The expression of wza, the first gene in the colanic acid production gene cluster, also increased. Real-time PCR analysis confirmed these results (Fig. 1A). Biochemical quantification revealed a higher level of colanic acid production by S3 mutants than by wild-type bacteria (Fig. 1B). When S3 mutant bacteria were observed under a transmission electron microscope, a thick capsule surrounding the cell was seen (Fig. 1C). Overproduction of colanic acid is known to affect the formation of mucoid colonies (17). Increased colanic acid production was seen in a group of randomly picked T7-resistant E. coli K-12 mutants (see Fig. S1 in the supplemental material). We further tested whether infection by other E. coli phages, T4 and lambda, also resulted in the appearance of colanic acid-overproducing resistant mutants of the K-12 strain. Increased colanic acid production was seen in a group of randomly picked T4-resistant or lambda-resistant E. coli K-12 mutants (see Fig. S1). Next, we investigated whether phage-resistant mutants with increased colanic acid production were also found among E. coli clinical isolates. Phage T7 could infect only 3 (F-939, F-722, F-521) of 140 clinical isolates. A group of T7-resistant mutants of F-939 were isolated and were found to overproduce colanic acid (see Fig. S2 in the supplemental material). Phage T4 could infect only 9 (F-521, F-508, F-519, F-730, F-810, F-859, F-873, F-915, and F-934) of 140 clinical isolates. A group of T4-resistant mutants of F-521 were isolated; they exhibited slight overproduction of colanic acid (see Fig. S2). Phage lambda could infect only 4 (F-939, F-521, F-705, and F-722) of the 140 clinical isolates. A group of lambda-resistant mutants of F-939 were isolated; they overproduced colanic acid (see Fig. S2). The appearance of colanic acid-overproducing phage-resistant E. coli mutants was less evident for clinical strains (especially T4-resistant bacteria) than for laboratory strains.
Gene expression in the S3 mutant relative to that in wild-type bacteria
Isolation of a T7-resistant mutant (S3 mt) overproducing colanic acid. (A) Total RNA from wild-type or mutant E. coli was isolated, and mRNA levels of rcsA, lon, and wza were measured using real-time reverse transcription-PCR. Glyceraldehyde-3-phosphate dehydrogenase was used as the internal control. Three independent experiments were performed. (B) l-Fucose from wild-type and mutant bacteria was quantified. Three independent experiments were performed. (C) Bacteria were stained with phosphotungstic acid and were observed under a transmission electron microscope. Bars, 0.5 μm.
The E. coli mutant S3 has a point mutation in the promoter region of the lon gene.After successive subcultures of resistant bacteria in the absence of T7, the colonies still appeared mucoid, suggesting that S3 harbors a genetic mutation. Thus, we checked the DNA sequences of genes involved in the Rcs phosphorelay system. No change was found in rcsA, rcsB, rcsC, rcsD, or the promoter of rcsA. However, we found a single base deletion in the −35 region in the promoter of the Lon protease gene (Fig. 2A). To determine whether this deletion actually changed promoter activity, a reporter plasmid expressing luciferase under the control of the wild-type or mutant promoter of the lon gene was constructed. A >50% decrease in activity from that of the wild-type promoter was observed for the mutant promoter (Fig. 2B). This suggests that the increased production of colanic acid by the T7-resistant mutant was due, at least in part, to this deletion mutation. Decreased expression of the lon gene led to an increase in the RcsA level, leading, in turn, to increased colanic acid production. When the wild-type lon gene under the control of the wild-type promoter was expressed in the S3 mutant, colanic acid production decreased to wild-type levels (Fig. 2C).
Point mutation in the lon gene promoter in S3 mutant bacteria. (A) The lon promoter regions from wild-type (wt) and mutant (S3 mt) bacteria were PCR amplified, and sequences were aligned (ClustalW). (B) Reporter plasmids containing the luciferase gene under the control of either the wild-type or the mutant lon promoter were constructed. Each plasmid was transformed into wild-type bacteria, and a luciferase assay was performed. Three independent experiments were performed. (C) The wild-type lon promoter and lon gene were cloned into the pBluescript vector and were used for the transformation of S3 mutant bacteria. Colanic acid from the wild type, the S3 mutant, and the transformant was quantified. Three independent experiments were performed.
Increased biofilm formation by the T7-resistant E. coli mutant.Since it has been reported that increased colanic acid production leads to increased biofilm formation (27), we tested biofilm formation by S3 mutant bacteria on a plastic surface. We observed a higher level of biofilm formation by the S3 mutant than by the wild type (Fig. 3A). In relation to this phenomenon, we tested whether the presence of phage T7 in the bacterial cultures altered biofilm formation. As shown in Fig. 3B, increased biofilm formation was observed when wild-type bacteria were cultured with phage T7 for 48 h, while a decrease in biofilm formation was observed after 8 h of culture. At the beginning of the culture, most bacteria would be infected and lysed. However, phage-resistant bacteria would appear almost instantly and would go on to dominate the culture. We analyzed colanic acid production in bacteria from the biofilm. Every bacterium we checked showed increased colanic acid production (Fig. 3C). The same phenomenon was observed for phages T4 and lambda (see Fig. S3 in the supplemental material).
Increase in biofilm formation by an E. coli phage-resistant mutant. (A) Wild-type or S3 mutant bacteria were cultured in 96-well polystyrene plates for 24 h. The culture medium was discarded, and the surface was stained with 0.4% crystal violet. (B) Wild-type bacteria were grown in 96-well polystyrene plates either in the presence or in the absence of phage T7. Biofilm was measured by staining with 0.4% crystal violet. (C) Wild-type bacteria were grown in the presence of phage T7 for 48 h. Biofilm was scraped, streaked onto an LB plate, and allowed to form colonies. Ten colonies were picked, and colanic acid from each colony was quantitated.
The S3 mutant inhibited phage adsorption.Since the S3 mutant produced colanic acid, it was possible that a capsular structure might have formed and inhibited phage adsorption. To test this hypothesis, we compared T7 adsorption to wild-type bacteria with that to the S3 mutant. For wild-type bacteria, 80% of phages were adsorbed to cells in 6 min, while no adsorption was observed for S3 mutant bacteria in the same period (Fig. 4A). To confirm this result, we isolated genomic DNA from phage T7 and used it for electroporation into the wild type or the S3 mutant. Infectious T7 particles were produced by both wild-type and S3 mutant bacteria (Fig. 4B). Thus, we concluded that the resistance of the S3 mutant to phage T7 infection was due to inhibition of phage adsorption to bacterial surfaces covered with colanic acid.
S3 mutant bacteria inhibited adsorption of phage T7. (A) Wild-type and S3 mutant bacteria were infected with phage T7 at an MOI of 0.1, and free phages in the culture were quantitated at the indicated time points. Three independent experiments were performed. (B) Genomic DNA of T7 was isolated and was used for electroporation into wild-type or S3 mutant bacteria. Cell growth was observed at 0 and 2 h postinfection in a spectrophotometer (600 nm). Plaques were observed on plates from lysates obtained after a 2-h incubation.
Infection of the S3 mutant by phage PBECO4.Due to the presence of the colanic acid capsule in the S3 mutant, phage T7 could not contact its receptor, resulting in resistance of the bacteria to the phage. Genomic analysis of phage PBECO4 has already been reported (21), and a putative gene encoding an enzyme with colanic acid-degrading activity was found. ORF547 (2,142 bp) of PBECO4 genomic DNA encodes a 78-kDa protein. In BLASTp analysis, the ORF547 product showed 57% similarity with the colanidase tailspike of phage phi92 (21). It was shown that PBECO4 successfully infected the S3 mutant, which was resistant to phage T7 (Fig. 5A). Next, we tested whether the protein product of ORF547 actually had colanic acid-degrading enzymatic (CAE) activity. The purified recombinant protein degraded colanic acid in both a dose-dependent and a time-dependent manner (Fig. 5B and C).
Bacteriophage PBECO4 infects E. coli mutant S3. (A) Growth of phage PBECO4 in the E. coli mutant S3 was observed. (B) The indicated amounts of purified recombinant CAE (product of PBECO4 ORF547) were added to cultures of S3 mutant bacteria. Colanic acid production from each culture was quantitated. (C) Purified recombinant CAE (product of PBECO4 ORF547) was added to cultures of S3 mutant bacteria for the indicated times. Colanic acid production from each culture was quantitated.
Treatment with a cocktail of phages T7 and PBECO4.Since the CAE activity of PBECO4 degraded the surface capsule formed by the E. coli S3 mutant, we explored whether this activity would help to expose receptors for T7, leading to successful infection of the resistant mutant host. Plaques produced by PBECO4 on the S3 mutant host were generally <1 mm in diameter. When the S3 mutant was coinfected with PBECO4 and T7, larger plaques appeared in addition to the regular small plaques produced by PBECO4 (Fig. 6A). Phages isolated from these larger plaques could infect wild-type E. coli but not the S3 mutant (data not shown). Thus, the larger plaques apparently resulted from infection by phage T7. We confirmed this by PCR (Fig. 6B). T7 DNA was detected in larger plaques. The growth of S3 mutant bacteria was inhibited more efficiently when the bacteria were coinfected with the two phages than when they were infected with PBECO4 alone (Fig. 6C). The E. coli S3 mutant remained resistant to infection by phage T7 alone. The same observation was made when the T7-resistant mutant of an E. coli clinical isolate (F-939) was infected with phages (Fig. 6D). In the presence of the purified recombinant CAE (ORF547 product) from phage PBECO4, T7 efficiently infected resistant mutant hosts (Fig. 6E). The cocktail effect was also observed for biofilm formation. When the S3 mutants were allowed to form biofilms for 48 h, addition of phage PBECO4 decreased biofilm formation by 50%, and the phage cocktail decreased biofilm formation by 80%, while the addition of T7 alone had little effect (Fig. 6F). When the wild-type bacteria were allowed to form biofilms for 48 h, biofilm formation increased 6-fold in the presence of phage T7, 5-fold in the presence of PBECO4, and 3.5-fold in the presence of the phage cocktail (Fig. 6G).
Cocktail effect of phages T7 and PBECO4. (A) S3 mutant bacteria were infected with either T7, PBECO4, or a cocktail (T7 plus PBECO4), and the resulting plaques were observed. Black arrows indicate typical (small) plaques of PBECO4; white arrows indicate larger plaques, not typical of PBECO4. (B) Genomic DNA was isolated from plaques of T7, PBECO4, or the larger-plaque phage in the cocktail treatment. PCR was performed with T7-specific primers. (C) Growth of the E. coli S3 mutant in the presence of either T7, PBECO4, or the phage cocktail. (D) Growth of the T7-resistant mutant of an E. coli clinical isolate (F-939) in the presence of either T7, PBECO4, or the phage cocktail. (E) Cocktail effect of purified recombinant colanic acid-degrading enzyme (CAE) and phage T7. The growth of the E. coli S3 mutant with the indicated amounts of CAE was observed. Three independent experiments were performed. (F) S3 mutant bacteria were incubated in a polystyrene 96-well plate in the presence of either SM buffer alone, T7, PBECO4, or a cocktail (T7 plus PBECO4) for 24 h. Biofilm was measured after staining with 0.4% crystal violet. (G) Wild-type bacteria were incubated in a polystyrene 96-well plate in the presence of either SM buffer alone, T7, PBECO4, or a cocktail (T7 plus PBECO4) for 48 h. Biofilm was measured after staining with 0.4% crystal violet.
DISCUSSION
The appearance of resistant bacteria upon treatment with bacteriophages is a problem in phage therapy (10). Many phages exploit bacterial surface virulence factors as their receptors, and phage-resistant mutant bacteria tend to lose these receptors, leading to a loss of bacterial virulence (28). However, the appearance of avirulent mutant bacteria is not disadvantageous to phage therapy. In this study, we demonstrated the appearance of phage-resistant bacteria surrounded by colanic acid capsules. Polysaccharide capsules confer increased virulence on bacteria (29), and the phage-resistant mutant studied here possesses a new phenotype that may be an unwanted side effect of phage therapy. Scholl et al. (30) have reported that the K1 capsule of E. coli is a barrier to phage T7 infection. The K1 capsule is distinct from the colanic acid capsule according to the classification of E. coli capsules by Whitfield and Roberts (31). The S3 mutant of E. coli, which was resistant to T7, harbored a base deletion in the promoter region of the lon gene. Decreased expression of Lon protease led to the activation of the Rcs phosphorelay system and the overproduction of colanic acid. There was an early report describing the selection of T7-resistant mutants with a mucoid colony morphology (32). The mutants resistant to T7 were also resistant to T3 and T4, which had receptors in the lipopolysaccharide layer. A non-9 mutant was selected and was shown to lose this resistance to T7. non-9 was mapped next to his and Su-1. From the complete genome of E. coli (GenBank accession number NC_000913), we can see that the wca genes are mapped to nucleotides 2113434 to 2133397, rcsA is mapped to nucleotides 2023968 to 2024591, lonA is mapped to nucleotides 458888 to 461242, the his genes are mapped to nucleotides 2089996 to 2426788, and Su-1 is mapped to nucleotides 2043468 to 2043557. Thus, it is probable that non-9 was a mutation in one of the wca genes. The appearance of a colanic acid-producing mutant was also observed in cultures of E. coli K 12 infected with phages T4 and lambda. Among E. coli clinical isolates, we could find strains infected by phage T7, T4, or lambda, while E. coli B strains are used for T7 infection in the laboratory. The appearance of a colanic acid-producing mutant was also observed with phage infection. Thus, the appearance of an encapsulated E. coli mutant after phage infection might be a common phenomenon.
The receptor for T7 is known to be lipopolysaccharide (33), and the capsular structure surrounding the bacteria should inhibit the phage-receptor interaction. Phage PBECO4 was recently isolated and characterized (21). Since one of the putative tail fiber genes (ORF547) harbored enzymatic activity degrading colanic acid (CAE), it was capable of infecting the colanic acid-overproducing S3 mutant of E. coli. Phage PBECO4 could also infect wild-type E. coli, suggesting that the phage receptor was a cell surface molecule other than colanic acid. In the presence of PBECO4 or the purified recombinant CAE, phage T7 successfully infected the S3 mutant. Many bacteria belonging to the Enterobacteriaceae produce colanic acid (16). Therefore, the utilization of this CAE or of CAE-producing phages could be an effective strategy for overcoming phage resistance. A similar strategy has been reported elsewhere (34). A T7 phage was engineered to express DspB, which could degrade EPS and could be used for effective removal of biofilms.
The colanic acid capsule is known to affect biofilm architecture and to allow for the formation of voluminous biofilms (35). From the data presented, it is evident that T7 affected biofilm formation biphasically. In the presence of phage T7, biofilm formation by wild-type E. coli decreased at first (at 8 h) but increased later (at 48 h). In a short time frame, T7 should lyse most wild-type bacteria, thus inhibiting the formation of biofilms. T7-resistant bacteria overproducing colanic acid should then appear and become dominant in the culture over a longer period. A thicker biofilm results. This is in accordance with a previous finding (36) that mucoid phage-resistant E. coli colonies appeared when a preformed biofilm was treated with phage T7 for 1 h. There is another report describing the successful suppression of E. coli biofilms by T7 infection (37). However, the authors did not observe biofilm formation long enough for the encapsulated mutants to become dominant; thus, they claimed that T7 was effective at inhibiting biofilm formation. As observed in our experiments, increased biofilm formation could be an unexpected result of phage therapy. Initial clearance of pathogenic bacteria was followed by the replacement of those bacteria by other, more virulent bacteria.
In summary, some of the phage-resistant bacteria appearing after phage therapy could be mutants with increased virulence (e.g., capsular mutants) (38–41). A phage cocktail is generally more effective than a single phage at suppressing the appearance of resistant mutants (42, 43). Usually, a phage cocktail consists of phages with different receptors on bacterial surfaces. However, in cases where a resistant mutant is encapsulated with polysaccharides, such a phage cocktail is of limited use. We have demonstrated that an advanced type of phage cocktail, containing a phage capable of degrading this capsule, is highly effective in controlling such a phage-resistant mutant.
ACKNOWLEDGMENTS
This work was supported by National Research Foundation of Korea grants NRF-2013R1A1A2060293, NRF-2010-0003280, and 2014 GRRC.
FOOTNOTES
- Received 12 August 2014.
- Accepted 14 November 2014.
- Accepted manuscript posted online 21 November 2014.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/AEM.02606-14.
- Copyright © 2015, American Society for Microbiology. All Rights Reserved.