Other infections which may be influenced by RNIs are those elicited by strains of uropathogenic Escherichia coli (UPEC). These bacteria are the primary causative agents of urinary tract infections, including both bladder and kidney infections (3). Within hours of initiation of a bladder infection, levels of nitrite in the urine increase threefold (24), and eventually gaseous levels of NO within the bladder increase 30- to 50-fold over uninfected controls (17). Much of this increase is likely due to the action of endothelial (e) NOS in the bladder, which has been shown to be upregulated and activated by intraperitoneal injection of E. coli lipopolysaccharide (14). The bacteria themselves, which can produce NO through nitrite reductases under conditions of low oxygen tension (7), may also contribute to the high levels of RNIs in an infected bladder. Like mycobacteria and several other pathogens that are known to be kept in check by RNIs, UPEC can persist long-term within a host as an intracellular pathogen and may lie in wait in quiescent reservoirs between recurrences of acute infection (19).

Considering the high levels of RNIs generated during the course of a UTI, we reasoned that UPEC, like many other pathogens, may have evolved mechanisms to resist the damaging effects of NO and its derivatives. Here we report on a UPEC isolate that has increased resistance to RNIs relative to a nonpathogenic K-12 reference strain. To identify factors involved in resistance of this UPEC isolate to nitrosative stress, we utilized a genetic selection and screening approach that identified a transcriptional regulator of polyamine synthesis. In addition, we demonstrate that polyamines are key factors in mediating RNI resistance and that production of the polyamine cadaverine is upregulated upon exposure to RNIs. To elucidate the mechanism of UPEC resistance to nitrosative stress, we test two mechanisms by which polyamines have been reported to protect against oxidative stress: chemical quenching and protection against mutagenesis.

Materials. Sodium nitrite, MES (morpholineethanesulfonic acid), polyamines, and antibiotics were obtained from Sigma-Aldrich (St. Louis, MO). IPTG (isopropyl-β-d-thiogalactopyranoside) was purchased from Teknova (Hollister, CA).

Strains and mutagenesis. MG1655 and UTI89 have been described previously (2, 19). UTI89 were transposon-mutagenized according to kit protocol (EZ-Tn5 <R6Kγori/KAN-2>Tnp Transposome Kit, Epicenter, Madison, WI). Targeted gene knockouts were created in UTI89 by Lambda Red-mediated linear transformation (8, 20) and verified by PCR of the affected region. Primer sequences are provided in Table 1.

TABLE 1. Primers used in this study

Growth conditions. All growth experiments were performed in 100 mM MES-buffered LB (pH 5), with or without added acidified sodium nitrite (ASN; sodium nitrite in 100 mM MES-buffered LB [pH 5]). Fresh MES-LB starter cultures were grown to stationary phase and then subcultured 1:100 into 6 ml of MES-LB in capped 20-by-150-mm borosilicate glass tubes. Cultures grew with shaking (225 rpm, tilted at a 30° angle) at 37°C, and optical densities at 600 nm (OD₆₀₀) were determined at the indicated times.

Identification of ASN-sensitive UTI89 mutants. Pools of transposon-mutagenized UTI89 mutants were cultured in 3 mM ASN (a concentration that is bacteriostatic for K-12 strains) for 1 h and then treated with 0.5 mg of penicillin G/ml to kill any dividing bacteria. Mutant pools were subjected to five rounds of this selection, and then screened individually for verification of sensitivity to ASN. Sites of transposon integration were identified by arbitrary PCR and sequencing as previously described (22).

Quantitation of cadaverine. Cadaverine was quantified by using a colorimetric assay as previously described (23, 25), with the following modifications. A total of 200 μl of total bacterial culture was thoroughly vortexed with 50 μl of chloroform and then mixed with 400 μl of 0.5 M K₂CO₃-5.1 mM 2′,4′,6′-trinitrobenzylsulfonic acid, and incubated for 5 min at 42°C. Then, 800 μl of toluene was used to extract the colored product, N,N′-bistrinitrophenylcadaverine, which was quantified by measuring the absorbance at 340 nm. Standard curves show that the assay is linear between cadaverine concentrations of 0.1 to 7 mM. When appropriate, titers of the cultures used in the cadaverine quantitation assays were determined to calculate the number of CFU present.

UTI89 is more resistant to ASN stress than MG1655. A well-established system for generating RNIs using ASN was used to examine bacterial RNI resistance in vitro (11, 18). When added to LB buffered to pH 5 with 100 mM MES, sodium nitrite dismutates to nitrous acid, which can then form nitric oxide and other RNIs (1). The pH of the media is not altered by addition of ASN. As shown in Fig. 1, the uropathogen UTI89 and the K-12 reference strain MG1655 grow similarly in unsupplemented MES-buffered LB (pH 5). However, addition of 3 mM ASN to the media reveals a differential phenotype: UTI89 is able to recover from ASN treatment far more readily than the nonpathogenic strain MG1655. In keeping with the reported bacteriostatic nature of acidified nitrite on E. coli (16), we found that MG1655 was not killed in ASN but merely unable to grow. Ongoing screens indicate that resistance to 3 mM ASN is not unique to UTI89 and is fairly common among UPEC isolates.

FIG. 1. The effect of 3 mM ASN on growth of E. coli strains UTI89 and MG1655. Fresh stationary phase starter cultures were subcultured 1:100 into 100 mM MES-buffered LB (pH 5) with or without added ASN. The cultures were grown with constant shaking at 37°C, (more ...)

ASN resistance in UPEC requires the cad gene cluster. To investigate the mechanism of UTI89 resistance to RNIs, we performed a genetic selection and screen for loss of the resistance phenotype. Pools of transposon-mutagenized UTI89 were cultured with ASN and penicillin. Because penicillin kills only dividing bacteria, this treatment selected for UPEC mutants which had lost resistance to ASN. Although this initial screen was not saturating, it yielded several mutants with reduced ASN resistance. One transposon insertion which produced sensitivity to ASN was in the gene cadC. CadC is the acid-inducible transcriptional activator of the cad operon, consisting of cadB, which encodes a lysine-cadaverine antiporter, and cadA, which encodes lysine decarboxylase (30). Plasmid-borne cadC (pPH2200), (30) complements the mutant phenotype, verifying the identity of the transposon mutation (Fig. 2A). Mutants with targeted deletion of cadA, cadB, or cadC all phenocopy the original transposon cadC mutation, demonstrating that the effects of the mutation are due to disruption of function of the cad operon and not some other unidentified target of the cadC transcriptional activator (Fig. 2B). Deletion of ldc, which encodes a second, more weakly expressed lysine decarboxylase in E. coli (15), did not affect the resistance of UTI89 to ASN. It should be noted that the actual time to recovery in ASN by wild-type UTI89 was variable between experiments, but the severely attenuated growth of MG1655 and the uncomplemented UTI89 cad mutants was invariant.

FIG. 2. CadC expression from pPH2200 complements the mutant phenotype of the cadC transposon mutant, and ΔcadA, -B, and -C UTI89 mutants are attenuated for growth in the presence of 3 mM ASN. (A) ODs were determined 24 h after subculture into media containing (more ...)

The lysine decarboxylase cadA converts lysine to cadaverine, a polyamine. Other major polyamines include putrescine, spermine, and spermidine. A succinct description of this class of nitrogenous bases was provided by Igarashi and Kashiwagi, who referred to polyamines as “mysterious modulators of cellular functions” (13). Polyamines are required for normal growth of all cell types tested thus far and can affect the function of proteins, DNA, and RNA, but the mechanisms behind most of their effects have not yet been fully elucidated. The most often cited role for cadaverine in bacteria is in acid tolerance. A low pH environment causes activation of the transcriptional regulator cadC (21) and the resultant lysine decarboxylase, cadA, consumes a proton while producing cadaverine. The antiporter cadB takes up lysine and excretes cadaverine, allowing the consumption of protons (and therefore modulation of pH within the microenvironment) to continue (26). In LB buffered with 100 mM MES at pH 5, the pH of wild-type and ΔcadA UTI89 cultures varied by at most 0.1 pH unit during the course of the experiments, casting doubt on the possibility that the pH modulating activity of the cad operon is crucial in this environment.

Exogenous polyamines promote UPEC growth in ASN. As shown in Fig. 3A, we found that addition of exogenous cadaverine, or other polyamines, can rescue the growth of ΔcadA (and other cad mutant) cultures in the presence of ASN. This contrasts against the role of the cad operon in acid tolerance, where the process of cadaverine synthesis—the consumption of a proton—is more important than the cadaverine itself. In control experiments, the addition of exogenous cadaverine in the absence of ASN had only modest effects on the growth of either wild-type UTI89 or the ΔcadA mutant (Fig. 3B). Our results echo those of Chattopadhyay et al. (5), which showed that polyamines, whether endogenous or exogenous, protect E. coli from the toxic effects of oxygen by unknown mechanisms.

FIG. 3. Rescue of the ΔcadA mutant by polyamines. (A) ODs of cultures with (+) or without (−) 3 mM ASN and with or without the indicated polyamines at 17 h after subculture with UTI89 or the isogenic ΔcadA mutant. This time point (more ...)

It is important to note that polyamines are apparently only one player in ASN stress resistance. As shown in Fig. 4A, the ASN-sensitive strain MG1655 also produces cadaverine, although only approximately half as much as UTI89. Virtually no cadaverine produced by either UTI89 or MG1655 was found associated with bacterial cell pellets, indicating that nearly all of the cadaverine produced by these strains is excreted. Addition of exogenous cadaverine (or spermine, spermidine, or putrescine) does enhance growth of MG1655 in ASN, but to a lesser degree than it affects the UTI89 cad mutants. When grown in 3 mM ASN and 3 mM cadaverine, MG1655 entered exponential phase at around the time that UTI89 ΔcadA entered stationary phase. Other as-yet-unidentified differences between MG1655 and UTI89, which share only ~56% of their protein coding sequences (C.-S. Hung and S. J. Hultgren, personal communication), undoubtedly also affect their differential growth phenotypes in ASN.

FIG. 4. Quantitation of cadaverine production. (A) Comparison of total cadaverine generated by MG1655, UTI89, and the ΔcadA UTI89 mutant grown for 3 h in unsupplemented MES-buffered LB (pH 5). (B) Time course of cadaverine produced per CFU of wild-type (more ...)

ASN treatment stimulates cadaverine production. Because exogenous cadaverine promotes growth in ASN, we wondered whether endogenous production of the polyamine is responsive to ASN stress. We found that exposure to 1 mM ASN (a molarity which slows, but does not halt, bacterial growth) does cause an increase in the amount of cadaverine produced per bacterial CFU in wild-type UTI89 (Fig. 4B). Within 3 h of subculture into ASN-containing media, there is a difference of >5-fold in the amount of cadaverine produced per CFU. We propose that this increase in production on a per-cell basis benefits all bacteria in the culture, since the majority of cadaverine produced appears to be secreted. Similar upregulation of cadaverine synthesis in response to 1 mM ASN is observed in the more ASN-sensitive strain MG1655, further emphasizing that factors other than cadaverine must also influence ASN resistance.

Cadaverine does not rescue by chemical quenching. One possibility for the mechanism of rescue by polyamines is simple quenching: RNIs react with amine groups, which are plentiful in polyamines, and so perhaps polyamines act as RNI scavengers. This hypothesis does have a precedent: the polyamine spermine has been shown to protect DNA from hydrogen peroxide-induced damage by acting directly as a free-radical scavenger (12). To determine whether cadaverine benefits bacterial growth in ASN by simply quenching the RNIs, the effects of coincubating cadaverine with ASN for 12 h prior to addition of the RNI-sensitive ΔcadA mutant UTI89 bacteria was tested as outlined in Fig. 5A. The cadaverine and ASN concentrations tested were those which have shown biologically relevant effects: 3 mM cadaverine rescues bacterial growth in 3 mM ASN (see Fig. 3A). Rather than enhancing growth of the ΔcadA mutant, as would be expected if RNI quenching occurred, prior coincubation of cadaverine with ASN instead actually reproducibly slowed bacterial growth slightly compared to samples in which ASN was not first coincubated with cadaverine (Fig. 5B). These data indicate that cadaverine does not facilitate bacterial growth in the presence of ASN by quenching RNIs. This conclusion is in keeping with our observation that MG1655 is able to produce substantial amounts of cadaverine but still remains sensitive to 3 mM ASN.

FIG. 5. Cadaverine does not rescue growth of ΔcadA UTI89 in ASN by quenching RNI activity. (A) Diagram of the experimental protocol used. The final concentrations of both ASN and cadaverine in the indicated cultures were 3 mM. (B) ΔcadA UTI89 (more ...)

Cadaverine does not decrease mutation frequency. Positively charged polyamines interact strongly with DNA and have been reported to affect the activity of many DNA-modifying enzymes such as polymerases and topoisomerases (6, 27). Given that RNIs can damage DNA (4, 10), we wondered whether polyamines may serve to reduce the frequency of ASN-induced mutation, as has been shown to occur with paraquat-induced oxidative stress (28, 29). This hypothesis was tested by measuring the frequency of mutation to rifamycin resistance (acquired by mutation of the rpoB gene) of wild-type and ΔcadA UTI89 growing in MES-LB (pH 5) with or without 1 mM ASN. This concentration of ASN does slow the growth of the ΔcadA mutant relative to the wild type, but both cultures undergo the same number of doublings and reach equivalent densities at stationary phase within the time course of the experiment. Although growth in ASN does increase the mutation frequency of both wild-type and ΔcadA bacteria beyond that observed without nitrosative stress, the results presented in Fig. 6 demonstrate that both strains experience the same level of mutagenesis. The inability to produce significant amounts of cadaverine did not notably affect mutation frequency. Similar results were obtained in assays for development of auxotrophic mutations (data not shown).

FIG. 6. Cadaverine does not abrogate the mutagenic effects of ASN. Wild-type UTI89 and the ΔcadA mutant were grown in MES-LB (pH 5) with or without 1 mM ASN for 24 h prior to plating on selective (plus 100 μg of rifamycin/ml) and nonselective (more ...)

Concluding remarks. One clear lesson from the current polyamine literature is that these small molecules can have maddeningly pleiotropic effects on biological functions. The work presented here represents the first demonstration that, among their various effects, polyamines are also key mediators of resistance to nitrosative stress in bacteria. Under acidic conditions, such as those used in these studies, cadC expression and consequently cadaverine synthesis are strongly induced (21). In contrast, none of the other polyamine biosynthetic pathways in E. coli, including those for spermidine and putrescine production, are acid inducible (6). As a result, cadaverine is likely the prominent polyamine generated by UTI89 in low-pH environments. This may explain why disruption of the cad gene cluster is sufficient to abrogate endogenous polyamine-dependent resistance to ASN and why exogenously added polyamines can overcome this defect. Interestingly, we have also shown that cadaverine production is actually amplified by exposure to ASN. How ASN stimulates cadaverine production is currently unknown. However, it is likely that this phenomenon further enhances UPEC survival in the face of RNIs generated during course of a UTI.

The mechanism(s) by which cadaverine or any polyamine mediates UPEC resistance to nitrosative stress remains unknown. However, in the present study we have ruled out two prime potential resistance mechanisms known to be applicable to oxidative stress, demonstrating that polyamines do not appreciably quench RNIs and that polyamines do not provide enhanced resistance to the mutagenic effects of RNIs. Whether or not cadaverine plays any role in modulating the expression of stress response or other genes in UPEC after exposure to RNIs remains unknown. As noted previously, the fact that the ASN-sensitive strain MG1655 does produce significant amounts of cadaverine indicates that factors other than polyamine production are at play. We are currently investigating what these other factors might be, as well as whether ASN resistance by UPEC isolates has any correlation to clinical outcome, including persistence and recurrence of the infection.

pPH2200 was made by Eric Olson and kindly provided by Jia Liu (Pfizer).

J.M.B. was supported by National Institutes of Health Genetics Training Grant T32-GM07464. This study was supported by grants DK068585 and DK069526 from the National Institutes of Health.