Abstract
We have screened the entire KEIO collection of 3,985 single-gene knockouts in Escherichia coli for increased susceptibility or resistance to the antibiotic bicyclomycin (BCM), a potent inhibitor of the transcription termination factor Rho. We also compared the results to those of a recent study we conducted with a large set of antibiotics (A. Liu et al., Antimicrob. Agents Chemother. 54:1393-1403, 2010). We find that deletions of many different types of genes increase sensitivity to BCM. Some of these are involved in multidrug sensitivity/resistance, whereas others are specific for BCM. Mutations in a number of DNA recombination and repair genes increase BCM sensitivity, indicating that DNA damage leading to single- and double-strand breaks is a downstream effect of Rho inhibition. MDS42, which is deleted for all cryptic prophages and insertion elements (G. Posfai et al., Science 312:1044-1046, 2006), or W3102 deleted for the rac prophage-encoded kil gene, are partially resistant to BCM (C. J. Cardinale et al., Science 230:935-938, 2008). Deletion of cryptic prophages also overcomes the increased BCM sensitivity in some but not all mutants examined here. Deletion of the hns gene renders the cell more sensitive to BCM even in the Δkil or MDS42 background. This suggests that BCM activates additional modes of cell death independent of Kil and that these could provide a target to potentiate BCM killing.
INTRODUCTION
In Escherichia coli, the ATP-dependent RNA-DNA helicase Rho (2, 31) terminates transcription at numerous sites (5, 10, 34). A recent study (4) revealed that critical Rho-dependent E. coli terminators are found in the cryptic prophages. Thus, Rho, aided by its cofactors NusA and NusG, is required to terminate transcription upstream of toxic foreign genes. Rho is the target for the antibiotic bicyclomycin (BCM) that binds specifically to Rho and inhibits its action (21, 39). BCM enhances transcription of horizontally transferred genes, including the various cryptic prophage genes that are otherwise silenced (4). One of these, the kil gene of the cryptic rac prophage, is lethal to the cell when expressed (18) and represents a major cause of cell death upon Rho inactivation. Cells lacking the kil gene are still sensitive to BCM but only at higher concentrations than that for the wild type (4).
What other pathways render cells sensitive to Rho inactivation? Previous work identified a number of recombinational repair functions that sensitized cells to rho mutations (17). We recently showed that Rho prevents replication fork collapse caused by transcription overrunning replication, resulting in DNA double-strand breaks (36). We wished to determine if other functions were involved in protecting the cell from BCM (intrinsic resistance). We therefore screened the KEIO collection (1) of 3,985 individual E. coli gene deletion mutants to identify mutants hypersensitive to BCM. All KEIO deletions were deliberately constructed so that the kan promoter is directed downstream and the kan gene is followed by no known terminator. This eliminates effects from transcriptional polarity, although translational polarity may, in some cases, still be possible. These mutants were compared with those in a similar study we carried out with 22 other antibiotics (23). We show that many different types of functions protect against Rho inactivation and that mutants lacking some of these functions sensitize the cell to BCM even in the absence of Kil function and other cryptic prophage genes.
MATERIALS AND METHODS
Synergy experiments.
BW25113 (11) was used in minimal medium A containing glucose (27). Approximately 105 cells from a growing culture were inoculated into 5 ml of medium and grown overnight. The following antibiotics and the corresponding target concentrations were used: bicyclomycin (BCM), 9 μg/ml; cephradine (RAD), 11.5 μg/ml; ciprofloxacin (CIP), 4 ng/ml; nitrofurantoin (NIT), 0.6 μg/ml; chloramphenicol (CHL), 0.5 μg/ml; tobramycin (TOB), 0.4 μg/ml; triclosan (TRI), 25 ng/ml; fusidic acid (FUS), 110 μg/ml; erythromycin (ERY), 5 μg/ml; and neomycin (NEO), 1.5 μg/ml. These target concentrations resulted in 60 to 80% growth compared with the growth in the absence of antibiotic in liquid culture. Four sets of triplicate cultures were prepared for each experiment: a minimal glucose medium control with no antibiotic, BCM alone, the antibiotic alone, and the combination of BCM with the antibiotic. The solvents for the stock solutions that were then diluted to prepare the final cultures were N,N-dimethylformamide for NIT and ethanol for CAM, ERY, and TRI. The control for these compounds also contained the same volumes of the respective solvent and the antibiotic used. Samples were analyzed using a Turner SP-830 spectrophotometer at a wavelength of 600 nm, using minimal medium A containing glucose as a blank. Samples with an optical density greater than 0.500 were diluted 1:4 in minimal glucose. Percent growth was determined by setting respective controls to 100% growth and determining the percent growth of the average of each set of triplicates. An interaction is defined as additive when the percent growth of the antibiotic pair is equal to the product of the two percent growth values for each antibiotic alone (37), here defined as the predicted additive effect (PAE). When the antibiotic pair has a percent growth significantly lower than the predictive additive effect, this is defined as a synergistic interaction.
High-throughput screening method.
For further details, see the report by Liu et al. (23). Briefly, cultures from the KEIO collection, maintained on 45 96-well microtiter plates, were taken out of frozen glycerol stored at −80°C using the Deutz cryoreplicator (14), transferred to sterile microtiter wells containing 0.5 ml of Luria broth (LB), and incubated at 37°C overnight. Approximately 3 to 5 μl of the resulting saturated culture was transferred to microtiter wells with fresh LB medium containing 50 μg/ml kanamycin to prevent the growth of contaminants (all strains taken from the KEIO collection are Kanr). After 3 h of growth, the Deutz cryoreplicator was used to print microdrops of the subculture onto LB agar plates (with no kanamycin) containing various concentrations of BCM and incubated at 37°C overnight.
Construction of double mutants.
The following strains (H. Mori and B. Wanner, unpublished data) were used as P1 donors to transduce mutants (e.g., the kil mutant) from the KEIO collection (1) or strain MDS42 (30) chloramphenicol resistance: the trxB::cat, fabF::cat, hns::cat, ydfA::cat, ydhT::cat, recA::cat, argO::cat, and dam:cat strains. Each strain contains a chloramphenicol-resistant marker that replaces the deleted gene region. The methodology used is as described by Miller (27).
Chemicals.
Chloramphenicol, nitrofurantoin, neomycin, tobramycin, erythromycin, fusidic acid, and triclosan were purchased from Sigma (St. Louis, MO). Ciprofloxacin was purchased from ICN Biomedicals, Inc. (Aurora, OH).
Screening for increased sensitivity to BCM.
The entire KEIO collection of almost 4,000 single-gene knockout mutants was screened using the high-throughput method described above against three subinhibitory concentrations of BCM: 20, 23, and 25 μg/ml. The MIC of BCM for the starting strain (BW25113) (11) under these conditions is 30 μg/ml. Approximately 4 ×105 exponentially growing cells were spotted onto LB agar plates containing either 20, 23, or 25 μg/ml BCM and allowed to grow overnight. Mutants displaying inhibited growth at any of the subinhibitory concentrations tested were determined to have increased susceptibility to BCM. All hypersensitive candidates were then repurified and retested using the same methodology.
Screening for increased resistance to BCM.
The entire KEIO collection was also screened for mutants with increased resistance to BCM compared to that of the wild type. The KEIO collection was screened against three higher levels of BCM: 40, 45, and 50 μg/ml. All resistant candidates were then purified from single colonies and retested using the same methodology.
RESULTS
Screening for BCM hypersensitivity.
The entire KEIO collection of 3,985 strains (1) was screened for mutants that are more sensitive than the wild type to 25 μg/ml BCM, using a Deutz cryoreplicator (14) and 96-well plates (see Materials and Methods for details). Mutants showing increased susceptibility were then purified and retested against a series of BCM concentrations, yielding a set of 76 strains with significantly enhanced susceptibility to BCM. Figure 1 displays examples of the data. Figure 2 shows all of the mutants with increased BCM sensitivity. Colors and numbers indicate the different functions lacking in each mutant (see the legend to Fig. 2). The strongest sensitivities are indicated in darker colors. Quantitative values are provided in Table 1. Whereas Rho is the primary target of BCM (39), many different functions provide intrinsic resistance to BCM, including those concerned with DNA replication, recombination, and repair, as well as functions involved in cell wall and cell membrane synthesis, chaperoning, protein synthesis, and general metabolism. In addition, several genes encoding transcriptional regulators suppress BCM sensitivity. Figure 2 also summarizes the susceptibilities of the BCM-sensitive mutants to a set of other antibiotics (23) to identify those with reduced intrinsic resistance to multiple drugs (e.g., the tolC, rimK, pgmB, and yciT mutants, among others) or those with a more specific pattern of resistance, such as trxA and trxB, which are hypersensitive only to BCM and rifampin (RIF). Ten of the 83 mutants (18%), shown in bold type, display enhanced susceptibility only to BCM.
Fig. 1.
Sensitivity analysis of bicyclomycin (BCM)-treated Escherichia coli gene knockout mutants. (A) Effects of 25 μg/ml BCM on E. coli growth. Ninety-six mutants from the KEIO collection were printed on LB agar plates with (right plate) and without (left plate) BCM. Two mutant strains on this plate, the fabF and trxB mutants, fail to grow on the plate containing 25 μg/ml BCM. (B) Differences in susceptibility to BCM among several single-gene knockout mutant strains compared against wild-type BW25113 E. coli strain.
Fig. 2.
Sensitivity profile for bicyclomycin (BCM). This sensitivity profile displays 76 single-gene knockout strains with increased sensitivity to BCM and their corresponding, collective sensitivity to 22 other antibiotics (MDS). These strains are organized by gene category, indicated by different colors: 1 (red), DNA replication, recombination, and repair; 1A (red), functions indirectly affecting category 1; 2 (green), transport, efflux, cell wall and cell membrane synthesis; 2A (teal), chaperones and functions related to category 2; 3 (orange), protein synthesis; 3A (orange), RNA processing; 4 (blue), central metabolic reactions; 5 (purple), regulation; 6 (yellow), prophage-encoded functions, cell adhesion; 7 (black), unassigned genes. Sensitivity to BCM is indicated with one of four levels of intensity: darker shades indicate stronger susceptibilities, lighter shades indicate medium susceptibilities, the lightest shades indicate medium-weak susceptibilities, and the lightest, halved shades indicate the weakest susceptibilities. The genes indicated in bold are uniquely sensitive to BCM. MDS indicates the degree of multidrug sensitivity of each BCM-sensitive strain. Darker brown shades indicate sensitivity to 8 or more other antibiotics, medium brown shades indicate sensitivity to 4 to 7 other antibiotics, light brown shades indicate sensitivity to 1 to 4 other antibiotics, and no shading indicates unique sensitivity to BCM. dsDNA, double-stranded DNA; ssDNA, single-stranded DNA; ACP, acyl carrier protein.
Table 1.
Bicyclomycin sensitivity of single mutants listed by level of susceptibilitya
| Strain | MIC (μg/ml) |
|---|---|
| BW25113 | 30 |
| fabF mutant | ≤20 |
| envC mutant | ≤20 |
| hns mutant | ≤20 |
| yciM mutant | ≤20 |
| yciB mutant | 20 |
| dnaK mutant | 20 |
| rimK mutant | 20 |
| ymfT mutant | 20 |
| rzpQ mutant | 20 |
| ydfA mutant | ≤23 |
| hscA mutant | ≤23 |
| rplA mutant | ≤23 |
| recO mutant | ≤23 |
| recA mutant | ≤23 |
| trxA mutant | ≤23 |
| trxB mutant | ≤23 |
| dapF mutant | ≤23 |
| yciT mutant | ≤23 |
| rlmE mutant | ≤23 |
| gpmM mutant | ≤23 |
| degP mutant | ≤23 |
| yebV mutant | ≤23 |
| minC mutant | 23 |
| surA mutant | 23 |
| lon mutant | 23 |
| rpsF mutant | 23 |
| pgpB mutant | 23 |
| rdgC mutant | 23 |
| ycjU mutant | 23 |
| ybaB mutant | 23 |
| csgG mutant | 23 |
| iscS mutant | 23 |
| ydhT mutant | 23 |
| tatC mutant | ≤25 |
| fur mutant | ≤25 |
| ydcX mutant | ≤25 |
| recJ mutant | ≤25 |
| uvrD mutant | ≤25 |
| pcnB mutant | ≤25 |
| tatB mutant | ≤25 |
| yfjY mutant | ≤25 |
| ydcS mutant | ≤25 |
| phoP mutant | ≤25 |
| argO mutant | 25 |
| tolC mutant | 25 |
| pal mutant | 25 |
| cedA mutant | 25 |
| hflK mutant | 25 |
| ygcO mutant | 25 |
| yebG mutant | 25 |
| ybjO mutant | 25 |
| rpmI mutant | 25 |
| rimI mutant | 25 |
| dam mutant | 25 |
| nlpC mutant | 25 |
| xseA mutant | >25 |
| ddlB mutant | >25 |
| recN mutant | >25 |
| nhaA mutant | >25 |
| pepN mutant | >25 |
| gntY mutant | >25 |
| rpmJ mutant | >25 |
| rppH mutant | >25 |
| cis mutant | >25 |
| bdm mutant | >25 |
| ydjO mutant | >25 |
| rpmG mutant | >25 |
| zwf mutant | >25 |
| dedD mutant | >25 |
| recB mutant | >25 |
| recC mutant | >25 |
| recF mutant | >25 |
| recG mutant | >25 |
| recR mutant | >25 |
| uvrC mutant | >25 |
| acrA mutant | >25 |
Gene knockouts are from the KEIO collection, using BW25113 as the starting strain. Strains listed in bold are uniquely sensitive to bicyclomycin.
Effect of deleting cryptic prophage genes.
BCM inhibition of Rho induces transcription read-through of many genes, including the kil gene of the cryptic rac prophage. Kil expression largely accounts for BCM sensitivity. Deletion of kil or of kil with all insertion elements and cryptic prophages (strain MDS42) yields significant resistance to BCM (4). We asked whether any of the BCM-sensitive mutants would be suppressed by deletions of cryptic prophages. Figure 3A confirms that deletion of kil increases resistance to BCM in otherwise wild-type E. coli. It also reveals that MDS42 is more BCM resistant than a strain deleted of kil alone. Figure 3B through D detail three mutations that increase wild-type E. coli susceptibility to BCM. With argO (Fig. 3B), both kil and the multiple deletions in MDS42 suppress the increased sensitivity and restore resistance to wild-type levels. On the other hand, hns strains remain sensitive to BCM in a kil or an MDS42 background (Fig. 3C). Interestingly, the BCM sensitivity of a recA strain is partially suppressed in an MDS42 background but not in a strain bearing only a kil deletion (Fig. 3D). We found that the recA mutant showed sensitivity in MG1655 (data not shown), which is isogenic with MDS42. Our interpretation is that the cryptic prophages contribute to the increased BCM sensitivity in recA strains.
Fig. 3.
BCM sensitivity of mutant strains lacking rac and other cryptic prophages. (A) Multiply deleted MDS42 strain and Δkil strain tolerate higher concentrations of BCM than the wild type. (B) MDS42 and Δkil background counteract increased sensitivity caused by deletion in argO and restore wild-type resistance to BCM. (C) Neither MDS42 nor Δkil background can counteract increased sensitivity to BCM caused by deletion of hns. (D) MDS42 background counteracts increased sensitivity to BCM caused by deletion of recA, but Δkil background does not.
Single-gene knockouts that increased resistance to BCM.
We also screened the entire KEIO collection for mutants more resistant to BCM than the wild type. Table 2 lists mutants that are resistant to BCM concentrations of 40 μg/ml or higher. Again, mutants defective in many types of functions are involved. Of particular note are mutants deleting the e14 prophage and the hfq gene (see Discussion). The cutoff here of 40 μg/ml is too high to detect the kil deletion in this collection, as this mutant does not grow above 32 μg/ml BCM (see Fig. 3).
Table 2.
Bicyclomycin-resistant single mutantsa
| Strain | MIC (μg/ml) | Gene production description |
|---|---|---|
| BW25113 | 30 | |
| hfq mutant | >50 | RNA chaperone facilitating sRNA-mRNA pairing interactions |
| acs mutant | 50 | Acetyl-coenzyme A synthetase (AMP forming) |
| proW mutant | 50 | Proline ABC transporter |
| rseB mutant | 50 | Anti-sigma factor |
| efeB mutant | 50 | Peroxidase; cryptic ferrous iron transporter |
| trmU mutant | 50 | tRNA methyltransferase |
| fliG mutant | 50 | Flagellar motor switch protein |
| flgI mutant | 50 | Flagellar P-ring protein |
| hinT mutant | 50 | Purine nucleoside phosphoramidase |
| ynfA mutant | 50 | Inner membrane protein |
| adhE mutant | 45 | Alcohol/acetaldehyde dehydrogenase |
| caiA mutant | 45 | Crotonobetainyl-coenzyme A reductase |
| ydjE mutant | 45 | MFS transporter |
| rhlE mutant | 45 | ATP-dependent RNA helicase |
| galE mutant | 45 | UDP-glucose 4-epimerase |
| yciU mutant | 45 | Predicted protein |
| yciS mutant | 45 | Conserved inner membrane protein |
| gloB mutant | 40 | Glyoxalase II |
| hybD mutant | 40 | Maturation peptidase for hydrogenase 2 |
| yneJ mutant | 40 | Predicted DNA-binding transcriptional regulator, LysR type |
| greA mutant | 40 | Transcription elongation factor |
| lysS mutant | 40 | Lysyl-tRNA synthetase |
| ydgA mutant | 40 | Conserved membrane protein |
| ygdD mutant | 40 | Conserved inner membrane protein |
| ymfJ mutant | 40 | e14 prophage |
| seqA mutant | 40 | Negative modulator of replication initiation |
| poxA mutant | 40 | Putative regulator of pyruvate oxidase |
Gene knockouts are from the KEIO collection, using BW25113 as the starting strain.
Pairwise interactions with other drugs.
We also examined a number of pairwise interactions as described by Yeh et al. (37). Table 3 shows the results of testing subinhibitory concentrations of BCM in minimal medium with subinhibitory concentrations of nine different antibiotics. Only the two aminoglycosides tested (NEO, TOB) were synergistic with BCM.
Table 3.
Pairwise interactions
| Antibiotic | % BCM-only growth | % antibiotic-only growth | % predicted additive growth (PAE) | % BCM + antibiotic growth | Interaction |
|---|---|---|---|---|---|
| Cephradine (RAD) | 70.9 | 65.1 | 46.2 | 55.0 | Additive |
| Ciprofloxacin (CIP) | 74.8 | 82.6 | 61.8 | 48.7 | Additive |
| Nitrofurantoin (NIT) | 71.2 | 78.3 | 55.7 | 54.9 | Additive |
| Chloramphenicol (CHL) | 70.0 | 64.8 | 45.4 | 45.3 | Additive |
| Triclosan (TRI) | 68.9 | 84.2 | 58.0 | 56.4 | Additive |
| Fusidic acid (FUS) | 69.5 | 81.5 | 56.6 | 61.4 | Additive |
| Erythromycin (ERY) | 80.5 | 75.4 | 60.7 | 62.3 | Additive |
| Neomycin (NEO) | 78.6 | 68.5 | 53.8 | 5.2 | Synergistic |
| Tobramycin (TOB) | 77.9 | 72.7 | 56.6 | 10.0 | Synergistic |
DISCUSSION
We employed the antibiotic bicyclomycin (BCM), a potent, specific inhibitor of Rho (39), to identify E. coli mutants with changes in sensitivity to the drug. Cardinale et al. (4) used microarrays to show that BCM treatment preferentially increases the expression of genes derived from recent horizontal transfer, including cryptic prophage as well as noncoding intergenic regions. A major cause of death upon Rho inactivation is expression of the kil gene of the rac cryptic prophage, whose product, KilR, inhibits the essential cell division function FtsZ (7). Cells deleted of kil were relatively resistant to BCM, despite massive gene disregulation due to read-through of terminators at the ends of operons. Our screen of a set of E. coli deletion mutants with altered BCM resistance (Fig. 1) defines the genes involved in intrinsic resistance to BCM and also provides information on the type of downstream events that occur after inactivation of Rho by BCM. Some of the gene deletions identified in Fig. 1 that increase sensitivity to BCM are involved in general intrinsic resistance to multiple antibiotics, as indicated by the intensity of color in the MDS (multidrug sensitivity) column. These include genes involved in recombination and recombinational repair of double-strand breaks (recA, -B, and -C), the main efflux pump in E. coli (acrA, tolC), genes involved in cell wall and cell membrane synthesis and integrity, and transporters and chaperones. This also includes several genes encoding ribosomal proteins (rplA, rrmJ) or their modification (rimK), a variety of transcriptional regulators (fur, hns, yciT), and genes with uncharacterized functions.
Recent work identified a number of recombinational repair functions that support resistance to BCM. It is proposed that Rho prevents replication fork collapse and double-strand breaks caused by replisome/RNA polymerase (RNAP) collisions (36). The numerous DNA recombination and repair functions in Fig. 1 support these findings. The recA, -B, and -C system repairs double-strand breaks, and the recFORJ system is involved in the repair of single-strand breaks (6, 24, 33). Mutants lacking the latter system increase sensitivity to only a select group of antibiotics, namely, those that damage DNA. Similarly, the recovery of uvrC, uvrD, and xseA in mutants (Fig. 1) further supports the idea that BCM treatment ultimately leads to DNA damage that is normally prevented by Rho. Interestingly, BCM does not induce the SOS system (36). We (36) have physically detected chromosomal breaks in cells treated with BCM. This underscores the complexities of antibiotic action (see, e.g., reference 19). Of particular interest are mutants that are specifically sensitive to BCM and not to any of the other antibiotics. Three of these appear to involve DNA repair proteins, the rdgC, yebG, and yfjY mutants. The ring-structured RdgC protein binds DNA and is associated with recombination and replication fork repair (3). Because it inhibits RecA-mediated DNA strand exchange reactions (15), it is considered to be a negative regulator of RecA action. Mahdi et al. (25) proposed that RecA drives the reversal of collapsed forks, rendering cells more dependent upon recombination to reestablish replication. In Fig. 1, we can see specific sensitivities within each category. For example, mutants lacking the thioredoxin/thioredoxin reductase system (the trxA and trxB mutants) (13) are specifically sensitive to BCM and RIF, and two mutants defective in septation, the minC and yciB mutants, are partially or completely specific for BCM. Mutants lacking the 50S ribosomal protein L36 (rpmI mutant) or the ribosomal protein alanine acetyltransferase (rimI mutant) are sensitive to only BCM among the antibiotics tested. Although many of the mutants defective in efflux pumps or membrane and cell wall synthesis and maintenance are multidrug sensitive, some members of this category are relatively specific, such as the argO mutant (strongly sensitive to BCM and CAM). Also, the protease-deficient lon mutant is hypersensitive only to BCM and metronidazole. This is consistent with the finding that Lon protease degrades SulA (28), a potent inhibitor of FtsZ (8, 20), reducing SulA's additive effect with BCM-stimulated expression of KilR. Highlighting the importance of FtsZ as a target of KilR is the finding that a deletion of hfq (22), a negative regulator of FtsZ (35, 38), yields cells that are more resistant to BCM (Table 2). For future applications, one might take advantage of these specific mutants in screening for antibiotics with properties similar to BCM.
Sozhamannan and Stitt (32) proposed a role for Rho in RNA degradation. Deletion of pcnB, encoding poly(A) polymerase I that polyadenylates RNAs, increased mRNA half–lives (29) and enhanced BCM sensitivity (Fig. 2). Moreover, deletion of rppH, which triggers mRNA degradation (12), also increased BCM sensitivity. Both of these results are consistent with the theory that Rho is involved in RNA turnover (32). It is possible that stabilization of the kilR transcript accounts for the hypersensitivity of these mutants, although an overall increase in RNA and RNA-DNA hybrids, toxic in Rho-deficient cells, is not ruled out.
We show that some of the increased susceptibilities to BCM could be overcome by deletion of kil alone or by deletion of all cryptic prophages and horizontally acquired genetic elements (strain MDS42). The deletions also increase BCM resistance in a wild-type background (4). Interestingly, in a BCM-sensitive mutant lacking the global regulator HNS, deletion of kil or placing the mutation in the MDS42 background did not enhance resistance. This raises the possibility of finding codrugs, or potentiators, of BCM by looking for inhibitors of, in this case, the HNS protein. Such an approach (9) might expand the use of BCM to other bacteria that lack the rac prophage. We (36) have studied a set of mutants, several of which are not in the KEIO collection, that also enhance the sensitivity of MDS42 to BCM. Figure 3 also shows that MDS42 is more resistant to BCM, in some circumstances, than strains lacking kil alone, suggesting that other kil-like-encoded functions missing in MDS42 contribute to BCM toxicity. A candidate for one of these functions might be the ymfJ gene of the e14 prophage (26), since deletion of this gene increases resistance to BCM (Table 2). This prophage carries a kil analogue, which Cardinale et al. (4) found to be upregulated approximately 10-fold by BCM.
We tested pairwise combinations of BCM with 9 different antibiotics (Table 3). Yeh et al. (37) have classified antibiotics into groups based on the patterns of their pairwise interactions. The data shown in Table 3, though not as extensive as that from Yeh et al., would at this point place BCM in the same grouping as CIP. This may not be surprising, given that BCM treatment results in expression of the Kil function that inhibits the cell division protein FtsZ and also inactivates Rho, leading to increased replisome/RNAP collisions and subsequent double-strand breaks (36), while CIP binds to DNA gyrase and blocks DNA synthesis, leading to double-strand breaks (16).
ACKNOWLEDGMENTS
We thank Barry Wanner for helpful advice.
This work was supported by grant GM37219 from the National Institutes of Health to M.E.G. and by grant ES0110875 to J.H.M.
Footnotes
Published ahead of print on 25 February 2011.
REFERENCES
- 1. Baba T., et al. 2006. Construction of Escherichia coli K-12 in-frame single-gene knockout mutants: the Keio collection. Mol. Syst. Biol. 2:2006.0008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Brennan C. A., Dombroski A. J., Platt T. 1987. Transcription termination factor Rho is an RNA-DNA helicase. Cell 48:945–952 [DOI] [PubMed] [Google Scholar]
- 3. Briggs G. S., et al. 2007. Ring structure of the Escherichia coli DNA-binding protein RdgC associated with recombination and replication fork repair. J. Biol. Chem. 282:12353–12357 [DOI] [PubMed] [Google Scholar]
- 4. Cardinale C. J., et al. 2008. Termination factor Rho and its cofactors NusA and NusG silence foreign DNA in E. coli. Science 230:935–938 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Ciampi M. S. 2006. Rho-dependent terminators and transcription termination. Microbiology 152:2515–2528 [DOI] [PubMed] [Google Scholar]
- 6. Clark A. J., Low K. B. 1988. Pathways and systems of homologous recombination in Escherichia coli, p. 155–215 In Low K. B. (ed.), The recombination of genetic material. Academic Press, San Diego, CA [Google Scholar]
- 7. Conter A., Bouché J.-P., Dassain M. 1996. Identification of a new inhibitor of essential division gene ftsZ as the kil gene of defective prophage Rac. J. Bacteriol. 178:5100–5104 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Cordell S. C., Robinson E. J., Lowe J. 2005. Inhibition of the SOS cell division inhibitor SulA and in complex with FtsZ. Proc. Nat. Acad. Sci. U. S. A. 100:7889–7894 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Cottarel G., Wierzbowski J. 2007. Combination drugs, an emerging option for antibacterial therapy. Trends Biotechnol. 25:547–555 [DOI] [PubMed] [Google Scholar]
- 10. Court D., et al. 1980. Control of transcription termination: a Rho-dependent termination site in bacteriophage lambda. J. Mol. Biol. 138:231–254 [DOI] [PubMed] [Google Scholar]
- 11. Datsenko K. A., Wanner B. L. 2000. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. U. S. A. 97:6640–6645 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Deana A., Celesnik H., Belasco J. G. 2008. The bacterial enzyme RppH triggers messenger RNA degradation by 5′ pyrophosphate removal. Nature 451:355–358 [DOI] [PubMed] [Google Scholar]
- 13. Debarbieux L., Beckwith J. 1998. The reductive enzyme thioredoxin 1 acts as an oxidant when it is exported to the Escherichia coli periplasm. Proc. Natl. Acad. Sci. U. S. A. 95:10751–10756 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Deutz W. A., et al. 2000. Methods for intense aeration, growth, storage, and replication of bacterial strains in microtiter plates. Appl. Environ. Microbiol. 66:2641–2646 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Drees J. C., Chitteni-Pattu S., McCaslin D. R., Inman R. B., Cox M. M. 2006. Inhibition of RecA protein function by the RdcG protein from Escherichia coli. J. Biol. Chem. 281:4708–4717 [DOI] [PubMed] [Google Scholar]
- 16. Drlica K., Zhao X. 1997. DNA gyrase, topoisomerase IV, and the 4-quinolones. Microbiol. Mol. Biol. Rev. 61:377–392 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Fassler J. S., Tessman I., Tessman E. 1985. Lethality of the double mutations rho rep and rho ssb in Escherichia coli. J. Bacteriol. 161:609–614 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Feinstein S. I., Low K. B. 1982. Zygotic induction of the rac locus can cause cell death in E. coli. Mol. Gen. Genet. 187:231–235 [DOI] [PubMed] [Google Scholar]
- 19. Hancock R. E. W. 2007. The complexities of antibiotic action. Mol. Syst. Biol. 3:142. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Huisman O., D'Ari R. 1981. An inducible DNA replication-cell division coupling mechanism in E. coli. Nature 290:797–799 [DOI] [PubMed] [Google Scholar]
- 21. Kohn H., Widger W. 2005. The molecular basis for the mode of action of bicyclomycin. Curr. Drug Targets Infect. Disord. 5:273–295 [DOI] [PubMed] [Google Scholar]
- 22. Lee T., Feig A. L. 2008. The RNA binding protein Hfq interacts specifically with tRNAs. RNA 14:514–523 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Liu A., et al. 2010. Antibiotic sensitivity profiles determined with an Escherichia coli gene knockout collection: generating an antibiotic bar code. Antimicrob. Agents Chemother. 54:1393–1403 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Lloyd R. G., Low K. B. 1996. Homologous recombination, p. 2236–2255 In Neidhardt F. C. (ed.), Escherichia coli and Salmonella: cellular and molecular biology. ASM Press, Washington, DC [Google Scholar]
- 25. Mahdi A. A., Buckman C., Harris L., Lloyd R. G. 2006. Rep and PriA helicase activities prevent RecA from provoking unnecessary recombination during replication fork repair. Genes Dev. 20:2135–2147 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Mehta P., Casjens S., Krishnaswamy S. 2004. Analysis of the lambdoid prophage element e14 in the E. coli genome. BMC Microbiol. 4:4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Miller J. H. 1992. A short course in bacterial genetics: a laboratory manual and handbook for Escherichia coli and related bacteria, p. 194–195 Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY [Google Scholar]
- 28. Mizusawa S., Gottesman S. 1983. The lon gene controls the stability of sulA protein. Proc. Natl. Acad. Sci. U. S. A. 80:358–362 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Mohanty B. K., Kushner S. R. 1999. Analysis of the function of Escherichia coli poly(A) polymerase I in RNA metabolism. Mol. Microbiol. 34:1094–1108 [DOI] [PubMed] [Google Scholar]
- 30. Posfai G., et al. 2006. Emergent properties of reduced genome Escherichia coli. Science 312:1044–1046 [DOI] [PubMed] [Google Scholar]
- 31. Roberts J. W. 1969. Termination factor for RNA synthesis. Nature 224:1168–1174 [DOI] [PubMed] [Google Scholar]
- 32. Sozhamannan S., Stitt B. L. 1997. Effects on mRNA degradation by Escherichia coli transcription termination factor Rho. J. Mol. Biol. 268:689–703 [DOI] [PubMed] [Google Scholar]
- 33. Spies M., Kowalczykowski S. C. 2005. Homologous recombination by RecBCD and RecF pathways, p. 389–403 In Higgins N. P. (ed.), The bacterial chromosome. ASM Press, Washington, DC [Google Scholar]
- 34. Stewart V., Landick R., Yanofsky C. 1986. Rho-dependent transcription termination in the tryptophanase operon leader region of Escherichia coli K-12. J. Bacteriol. 166:217–223 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Takada A., Wachi M., Nagai K. 2009. Negative regulatory role of the Escherichia coli hfq gene in cell division. Biochem. Biophys. Rec. Commun. 266:579–583 [DOI] [PubMed] [Google Scholar]
- 36. Washburn R. S., Gottesman M. E. 2011. Transcription termination maintains chromosome integrity. Proc. Natl. Acad. Sci. U. S. A. 108(2):792–797 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Yeh P., Tschumi A. I., Kishony R. 2006. Functional classification of drugs by properties of their pairwise interactions. Nat. Genet. 28:489–494 [DOI] [PubMed] [Google Scholar]
- 38. Zambrano N., et al. 2009. Involvement of HFq protein in the post-transcriptional regulation of E. coli bacterial cytoskeleton and cell division proteins. Cell Cycle 8:2470–2472 [DOI] [PubMed] [Google Scholar]
- 39. Zwiefka A., Kohn H., Widger W. R. 1993. Transcription termination factor Rho: the site of bicyclomycin inhibition in Escherichia coli. Biochemistry 32:3564–3570 [DOI] [PubMed] [Google Scholar]