• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of plospathPLoS PathogensSubmit to PLoSGet E-mail AlertsContact UsPublic Library of Science (PLoS)View this Article
PLoS Pathog. May 2009; 5(5): e1000448.
Published online May 29, 2009. doi:  10.1371/journal.ppat.1000448
PMCID: PMC2680622

Fitness of Escherichia coli during Urinary Tract Infection Requires Gluconeogenesis and the TCA Cycle

Jorge E. Galán, Editor

Abstract

Microbial pathogenesis studies traditionally encompass dissection of virulence properties such as the bacterium's ability to elaborate toxins, adhere to and invade host cells, cause tissue damage, or otherwise disrupt normal host immune and cellular functions. In contrast, bacterial metabolism during infection has only been recently appreciated to contribute to persistence as much as their virulence properties. In this study, we used comparative proteomics to investigate the expression of uropathogenic Escherichia coli (UPEC) cytoplasmic proteins during growth in the urinary tract environment and systematic disruption of central metabolic pathways to better understand bacterial metabolism during infection. Using two-dimensional fluorescence difference in gel electrophoresis (2D-DIGE) and tandem mass spectrometry, it was found that UPEC differentially expresses 84 cytoplasmic proteins between growth in LB medium and growth in human urine (P<0.005). Proteins induced during growth in urine included those involved in the import of short peptides and enzymes required for the transport and catabolism of sialic acid, gluconate, and the pentose sugars xylose and arabinose. Proteins required for the biosynthesis of arginine and serine along with the enzyme agmatinase that is used to produce the polyamine putrescine were also up-regulated in urine. To complement these data, we constructed mutants in these genes and created mutants defective in each central metabolic pathway and tested the relative fitness of these UPEC mutants in vivo in an infection model. Import of peptides, gluconeogenesis, and the tricarboxylic acid cycle are required for E. coli fitness during urinary tract infection while glycolysis, both the non-oxidative and oxidative branches of the pentose phosphate pathway, and the Entner-Doudoroff pathway were dispensable in vivo. These findings suggest that peptides and amino acids are the primary carbon source for E. coli during infection of the urinary tract. Because anaplerosis, or using central pathways to replenish metabolic intermediates, is required for UPEC fitness in vivo, we propose that central metabolic pathways of bacteria could be considered critical components of virulence for pathogenic microbes.

Author Summary

Bacteria that cause infections often have genes known as virulence factors that are required for bacteria to cause disease. Studying virulence factors such as toxins, adhesins, and secretion and iron-acquisition systems is a fundamental part of understanding infectious disease mechanisms. In contrast, little is known about the contribution of bacterial metabolism to infectious disease. This study shows that E. coli, which cause most urinary tract infections, utilize peptides as a preferred carbon source in vivo and requires some, but not all, of the central metabolic pathways to infect the urinary tract. Specifically, pathways that can be used to replenish metabolites, known as anaplerotic reactions, are important for uropathogenic E. coli infections. These findings help explain how metabolism can contribute to the ability of bacteria to cause a common infection.

Introduction

Traditional studies of bacterial pathogenesis have focused on pathogen-specific virulence properties including toxins, adhesins, secretion, and iron acquisition systems, and mechanisms to avoid the innate and adaptive immune response. Examining bacterial metabolism during the course of an infection is also critical to further our understanding of pathogenesis and identifying potential targets for new antimicrobial agents. Infectious diseases represent a serious threat to global health because many bacteria that cause disease in humans such as Staphylococcus aureus, Mycobacterium tuberculosis, and E. coli are steadily developing resistance to many of the available treatments [1][3]. Since the introduction of antibiotics in the last century, the emergence of bacteria that resist these compounds has rapidly outpaced the discovery and development of new antimicrobial agents [4]. The need to understand bacterial physiology during infection of the host is critical for the development of new antimicrobials or antibiotics that will reduce their burden upon human health.

Among common infections, urinary tract infections (UTI) are the most frequently diagnosed urologic disease. The majority of UTIs are caused by E. coli and these uropathogenic E. coli (UPEC) infections place a significant financial burden on the healthcare system by generating annual costs in excess of two billion dollars [5],[6]. Because UTIs are a significant healthcare burden and E. coli is one of the best studied model organisms for studying metabolism, these traits can be exploited to understand and identify metabolic pathways that are required for the growth of the bacterium during infection of the host.

Despite being arguably the most studied organism, E. coli metabolism during colonization of the intestine has only recently been explored [7],[8]. Commensal E. coli acquires nutrients from intestinal mucus, a complex mixture of glycoconjugates, and subsequently expresses genes involved in the catabolism of N-acetylglucosamine, sialic acid, glucosamine, gluconate, arabinose and fucose [8],[9]. E. coli mutants in the Entner-Doudoroff and glycolytic central metabolic pathways have diminished colonization levels reflecting the importance of sugar acid catabolism [8]. These findings suggest that commensal E. coli uses multiple limiting sugars for growth in the intestine [8]. Together, this developing body of evidence supports the assertion that E. coli grows in the intestine using simple sugars released by the breakdown of complex polysaccharides by anaerobes [9],[10].

Much less is known about the metabolism of enteric pathogens during colonization of the gastrointestinal tract. Enterohemorrhagic E. coli (EHEC) O157[ratio]H7 requires similar carbon metabolic pathways as do commensal strains, however, mutations in pathways that utilize galactose, hexuronates, mannose, and ribose resulted in colonization defects only for EHEC [9]. It was also found that multiple mutations in a single EHEC strain had an additive effect on colonization levels suggesting that this pathogen depends on the simultaneous metabolism of up to six sugars to support the colonization of the intestine [9]. When faced with limiting sugars due to consumption by other colonizing bacteria, EHEC may switch from glycolytic to gluconeogenic substrates to sustain growth in the intestine [11]. Synthesis and degradation of glycogen, an endogenous glucose polymer, plays an important role for EHEC and pathogenic Salmonella during colonization of the mouse intestine presumably by functioning as an internal carbon source during nutrient limitation [12][14]. Although it is not known which external carbon sources are used by S. enterica serovar Typhimurium during colonization it has been demonstrated that full virulence requires the conversion of succinate to fumarate in the tricarboxylic acid (TCA) cycle [15],[16]. These studies have contributed much to the understanding of the in vivo metabolic requirements of EHEC colonization; however, these studies were done in an animal model that is not suitable for studying pathogenesis because these animals do not exhibit signs of EHEC infection [9],[11],[13].

In contrast to the nutritionally diverse intestine, the urinary tract is a high-osmolarity, moderately oxygenated, iron-limited environment that contains mostly amino acids and small peptides [17],[18]. The available studies on UPEC metabolism during UTI has revealed that the ability to catabolize the amino acid D-serine in urine, which not only supports UPEC growth, appears important as a signaling mechanism to trigger virulence gene expression [19],[20]. Metabolism of nucleobases has been demonstrated to play a role for UPEC colonization of the urinary tract; signature-tagged mutagenesis screening identified a mutant in the dihydroorotate dehydrogenase gene pyrD that was outcompeted by wild-type UPEC in vivo [21] and in a separate transposon screen a gene involved in guanine biosynthesis, guaA, was identified and found to be attenuated during experimental UTI [22].

To better understand bacterial metabolism during infection, we used a combination of comparative proteomics and systematic disruption of central metabolism to identify pathways that are required for UPEC fitness in vivo. By examining the expression of UPEC cytoplasmic proteins during growth in human urine, we confirmed that E. coli is scavenging amino acids and peptides and found that disruption of peptide import in UPEC significantly compromised fitness during infection. Consistent with the notion that peptides are a key in vivo carbon source for UPEC, only mutations ablating gluconeogenesis and the TCA cycle demonstrated reduced fitness in vivo during experimental UTI. These findings represent the first study of pathogenic E. coli central metabolism in an infection model and further our understanding of the role of metabolism in bacterial pathogenesis.

Results

Proteomic profile for uropathogenic E. coli growing in urine

Culturing UPEC in human urine partially mimics the urinary tract environment and has proven to be a useful tool to identify bacterial genes and proteins involved in UTI [18], [22][24]. Because it is well established that urine is iron-limited and our previous studies clearly demonstrated that the majority of differentially expressed genes and proteins are involved in iron acquisition [18],[23], we determined the protein expression profile of E. coli CFT073 during growth in human urine and compared that with bacterial cells cultured in iron-limited LB medium to unmask proteins involved in processes other than iron metabolism. Using this strategy and 2D-DIGE it was possible to visualize 700 cytoplasmic protein spots, 84 of which were differentially expressed (P<0.05) between urine and iron-limited LB medium (Fig. 1). Of these, 56 were more highly expressed in human urine (green) than in iron-limited LB medium, while 28 demonstrated greater expression in iron-limited LB medium (red) than in urine (Fig. 1).

Figure 1
Fluorescence difference in gel electrophoresis (2D-DIGE) of UPEC cytoplasmic proteins during growth in urine.

Proteins induced in human urine with >2-fold differences from expression levels in iron-limited LB medium were identified by tandem mass spectroscopy (Table 1). The results indicate that E. coli growing in urine are expressing proteins involved in the catabolism of pentose sugars; XylA (xylose isomerase), AraF (high-affinity arabinose-binding protein), and the non-oxidative pentose phosphate pathway enzyme TalA (transaldolase) were induced 5.25-, 2.02-, and 5.66-fold, respectively (P<0.001) (Table 1). Other proteins that were induced are the involved in metabolism of the sugar acids gluconate (UxuA, mannonate dehydratase), gluconolactone (YbhE, 6-phosphogluconolactonase), sialic acid (NanA, N-acetylneuraminate lyase), and fructose (FruB, fructose-specific IIA/FPr PTS system component). Multiple isoforms of the periplasmic dipeptide and oligopeptide substrate-binding proteins DppA and OppA were also induced (>2-fold, P<0.009) in urine confirming the notion that amino acids and small peptides are being acquired from this milieu (Table 1). Proteins involved in amino acid metabolism were also identified and include SerA (D-3-phosphoglycerate dehydrogenase) that is involved in serine biosynthesis and two enzymes in the arginine biosynthesis pathway, ArgG (argininosuccinate dehydrogenase) and SpeB (agmatinase) (Table 1). As expected, none of the proteins identified were involved in iron uptake or metabolism, although DppA has been reported to bind heme albeit with less affinity than dipeptide substrates [25].

Table 1
UPEC cytoplasmic proteins differentially expressed in human urine.

Notably, there was an increase in abundance for two central metabolism enzymes, TalA, as mentioned above, and TpiA that was increased 4.58-fold (P<0.0001) in urine (Table 1). TalA, a non-oxidative pentose phosphate pathway enzyme, converts sedoheptulose-7-phosphate and glyceraldehyde-3-phosphate to erythrose-4-phosphate and fructose-6-phosphate. Due to the transfer of the glycolytic intermediate glyceraldehyde-3-phosphate by TalA, this enzyme is an important link between the pentose phosphate pathway and glycolysis [26]. TpiA is a glycolytic enzyme that catalyzes the reversible isomerization of glyceraldehyde-3-phosphate and dihydroxyacetone phosphate [27]. The induction of TalA and TpiA suggested that the coupling of the pentose phosphate pathway and glycolysis or gluconeogenesis via the transfer and isomerization of glyceraldehyde-3-phosphate may be an important route of carbon flux through these central pathways during the bacterium's growth in human urine.

Contribution of genes induced in urine to UPEC fitness in vivo

To determine whether some proteins identified by 2D-DIGE are required for UPEC fitness during UTI, CFT073 mutants were constructed in the genes: talA, xylA, tpiA, serA, speB, uxuA, nanA, argG, araF, dppA, and oppA. For these studies, an experimental competition between each mutant strain and wild-type parental CFT073 was performed. Wild-type UPEC and the mutant strain were prepared in a 1[ratio]1 ratio and transurethrally inoculated into the bladders of mice. The number of mutant (kanamycin-resistant) and wild-type (kanamycin-sensitive) bacteria recovered from the bladder and kidneys was determined by plating the tissue homogenates for CFU on both LB agar and LB agar containing kanamycin. Mutants containing defects in genes that affect fitness in vivo are out-competed by the wild-type strain when inoculated into the same animal. This was determined by comparing the ratio of colony forming units (CFU) of bacteria recovered from the infection to the ratio of bacteria contained within the inoculum to obtain a competitive index (CI). A CI>1 indicates the wild-type out-competes the mutant strain and a CI<1 indicates the wild-type is out-competed by the mutant. In these series of experimental infections, only mutants defective in peptide transport (ΔdppA and ΔoppA) were dramatically out-competed by wild-type UPEC in vivo, CI>50, P<0.005 for the bladder (Table 2). One additional mutant, ΔtpiA, that functions in both glycolysis and gluconeogenesis, was out-competed by wild-type in the kidneys at 48 hpi, CI = 2.54, P = 0.0206 (Table 2).

Table 2
In vivo fitness for select 2D-DIGE mutants.

Despite the lack of attenuation in vivo for the many of the mutants, these results reveal a number of important findings. The agmatinase mutant ΔspeB out-competed wild-type in the bladder at 48 hpi, CI = 0.14, P = 0.0122 (Table 2). Agmatinase is part of arginine metabolism and catalyzes the formation of the polyamine putrescine and urea from agmatine and H2O. This suggests that accumulation of agmatine or reduced production of urea and putrescine by the mutant may provide a modest advantage over wild-type UPEC during infection of the bladder. CFT073 ΔargG was unable to grow in MOPS defined medium unless supplemented with 10 mM arginine (Fig. 2A), validating the expected auxotrophic phenotype. Similarly, the ΔserA serine auxotroph required supplementation with either 10 mM serine or glycine in MOPS, D-serine was unable to rescue the in vitro growth defect (Fig. 2B). Lack of arginine or serine biosynthesis had little effect upon the ability of UPEC to grow logarithmically in human urine, although the ΔargG mutant consistently entered stationary phase at a lower cell density, with an O.D.600 of 0.45±0.04 compared to 0.59±0.03 for wild-type (P = 0.051) (Fig. 2C). When tested for in vivo fitness, neither the ΔargG nor ΔserA strain were significantly out-competed by wild-type UPEC at 48 hpi (Fig. 2D, 2E, and Table 2). Additionally, there was no preference for serine over arginine or vice versa for UPEC colonization at 48 hpi. When the auxotrophic strains were co-inoculated into the same mice both mutants were recovered at similar levels (Fig. 2F). These data clearly demonstrate that there are sufficient concentrations of arginine, serine and/or glycine in the urinary tract to support growth of these auxotrophic strains.

Figure 2
In vivo contribution of UPEC arginine and serine biosynthesis.

As mentioned, deletion of the genes encoding periplasmic peptide substrate-binding proteins, dppA and oppA, had the greatest impact on UPEC fitness in vivo of the CFT073 mutants in genes whose products were induced during growth in human urine (Table 2). The dipeptide transport mutant, ΔdppA, failed to maintain colonization in the bladder at 48 hpi, 11/11 bladders had undetectable levels (<200 CFU/g) for this mutant, while wild-type levels from the same bladders reached a median of 104 CFU/g (P = 0.0020) (Fig. 3A). Because these mice had low levels of recoverable UPEC from the kidneys it was not possible to determine the contribution of dipeptide transport for kidney colonization. Import of oligopeptides via the OppA substrate-binding protein is also required for UPEC fitness in vivo. CFT073 ΔoppA was out-competed nearly 500[ratio]1 wild-type[ratio]mutant in the bladder (Table 2) with a 3-log reduction in the median CFU/g from bladder tissue at 48 hpi (P = 0.0047) (Fig. 3B). In these co-challenge infections, wild-type UPEC colonized 10/16 (62%) of kidneys, while ΔoppA was detectable in 4/16 (25%) of kidneys at 48 hpi. The ratio of wild-type[ratio]mutant recovered from the kidneys at this time point was 156[ratio]1 (Table 2) where wild-type UPEC had 3-logs greater CFU/g than ΔoppA (P = 0.0420) (Fig. 3B). Together, the in vivo fitness defect for CFT073 harboring a deletion of either dppA or oppA suggests that peptides may be an important carbon source for UPEC during urinary tract infection.

Figure 3
In vivo contribution of UPEC peptide substrate-binding proteins.

Previously, we have shown that the low copy pGEN plasmid is maintained in CFT073 in the absence of antibiotic pressure for up to 48 h [28]. Using this ampicillin resistant plasmid system, we cloned the entire dppA gene including 200 bp upstream from the predicted start site of translation and introduced the resulting construct, pGEN-dppA, into the CFT073 ΔdppA strain. To determine if it was possible to complement the ΔdppA defect in vivo, co-challenge infections were performed as described and modified to enumerate bacteria in tissue homogenates by plating on agar containing ampicillin (wild-type CFT073 harboring pGEN) or ampicillin and kanamycin (CFT073 ΔdppA containing pGEN or pGEN-dppA). The ΔdppA mutant containing empty vector (pGEN-) demonstrated the expected fitness defect in bladder colonization when co-inoculated with wild-type CFT073 (pGEN-) (P = 0.0002) while ΔdppA containing a wild-type copy of dppA (pGEN-dppA) restored colonization to wild-type levels in the bladder at 48 hpi (Fig. 3C). Although both mutant (pGEN-) and wild-type (pGEN-) demonstrated poor colonization in the kidneys of these animals, complementation of ΔdppA (pGEN-dppA) resulted in a 2-log increase in median kidney CFU/g at 48 hpi (Fig. 3D).

Fitness of UPEC central carbon metabolism mutants during UTI

The requirement for peptide transport for UPEC fitness during infection implicates peptides as an important carbon source in vivo. This predicts that certain central metabolism pathways that operate during catabolism of amino acids or peptides may be more important for in vivo growth of UPEC than pathways that function primarily to catabolize sugars. To test the role of central metabolic pathways during an actual infection mutants were constructed in UPEC strain CFT073 to produce defects in glycolysis (pgi, phosphoglucose isomerase and tpiA, triosephosphate isomerase) [29], the Entner-Doudoroff pathway (edd, 6-phosphogluconate dehydratase) [10], the oxidative branch (gnd, 6-phosphogluconate dehydrogenase) and the non-oxidative branch (talA, transaldolase) of the pentose phosphate pathway [26], gluconeogenesis (pckA, phosphoenolpyruvate carboxykinase) [30], and the TCA cycle (sdhB, succinate dehydrogenase) [31]. The in vitro growth of these central metabolism mutants were examined and compared to wild-type UPEC during culture in human urine, LB medium, and MOPS defined medium containing 0.02% glucose. All of the central metabolism mutants produced similar logarithmic growth as wild-type when cultured in human urine (Fig. 4A) and LB medium (data not shown) under defined inoculation conditions. As expected, only mutants with defects in glycolysis demonstrated diminished growth in MOPS medium containing glucose as the sole carbon source (Fig. 4B). The Δpgi strain produced an extended lag phase of 5.5±1.1 h compared with wild-type (P = 0.001) and ΔtpiA failed to reach exponential phase after 18 h (Fig. 4B). These data and the indistinguishable growth of the glycolysis mutants from wild-type in urine supported the proteomics data and indicated that UPEC growing in urine utilizes carbon sources other than glucose.

Figure 4
In vitro growth of UPEC central metabolism mutants.

To determine the role for central metabolism during E. coli infection of the urinary tract, the ascending model of murine UTI was used as described above to measure the impact that a lesion in central metabolism has upon the relative fitness of the strain in vivo. Mutants with defects in glycolysis had levels of colonization in the bladder at 48 hpi similar to wild-type (P>0.400) (Fig. 5A and 5B). In the kidneys, Δpgi CFU/g were comparable to wild-type (Fig. 5A), while ΔtpiA demonstrated a 10-fold reduction in the median CFU/g (P = 0.0206) (Fig. 5B). The pentose phosphate pathway mutants, Δgnd (Fig. 5C) and ΔtalA (Table 2), were not significantly out-competed by wild-type in vivo. The mutant with a defect in the Entner-Doudoroff pathway (Δedd) also was not impaired in the ability to infect both the bladder and kidneys as indicated by its similar colonization to wild-type at 48 hpi (Fig. 5D). UPEC in vivo fitness was significantly reduced in the TCA cycle mutant ΔsdhB, this mutation resulted in a 50-fold reduction in median CFU/g in the bladder (P = 0.0134) and a 1.5-log decrease in kidney CFU at 48 hpi (P = 0.0400) (Fig. 5E). This defect in the TCA cycle impacted fitness to a greater extent in the bladder, where 11/15 (73%) of mice had undetectable levels of mutant bacteria, than in the kidneys where 6/15 (40%) mice had undetectable counts (Fig. 5E). The gluconeogenesis mutant, ΔpckA had a 2-log reduction in median CFU/g in both the bladder (P = 0.0005) and kidneys (P = 0.0322) and half of the mice (7/14) displayed undetectable levels of ΔpckA at 48 hpi (Fig. 5F).

Figure 5
In vivo fitness of UPEC central metabolism mutants.

To verify that this mutation is non-polar as expected and the defect in colonization is not due to a secondary mutation, in vivo complementation experiments were conducted. The ΔpckA mutant with the pGEN empty vector demonstrated a 2-log reduction in CFU/g at 48 hpi (P = 0.0039) in the bladder when co-inoculated into mice with wild-type UPEC containing pGEN (Fig. 6). When CFT073 ΔpckA (pGEN-pckA) were co-inoculated with CFT073 (pGEN-) there was no significant difference in bladder CFU/g at 48 hpi between the strains (Fig. 6). Thus, by re-introducing the pckA gene into the mutant it was possible to complement the ΔpckA defect in bladder colonization at 48 hpi.

Figure 6
In vivo complementation of UPEC ΔpckA.

The in vitro growth and in vivo fitness for the UPEC central metabolism mutants is summarized in Table 3. As expected, only mutations in glycolysis had a negative effect on growth in defined medium with glucose. Only gluconeogenesis or TCA cycle mutants demonstrated reduced persistence at 48 hpi in both the bladder and kidneys (Table 3). Non-oxidative and oxidative pentose phosphate pathway and Entner-Doudoroff pathway mutants did not demonstrate any colonization defect and of the glycolytic mutants only the triosephosphate isomerase deletion had a measurable defect in the kidneys but not in the bladder (Table 3). Together, the fitness defect for the peptide transport mutants and these data indicate UPEC could be using amino acids as the primary carbon source during infection. Surprisingly, there was no correlation between the ability of the central metabolism mutants to grow in human urine ex vivo and grow in the urinary tract in vivo.

Table 3
Growth of central metabolism mutants in vitro and in vivo.

Discussion

Bacterial pathogenesis traditionally involves studying virulence traits involved in the production of toxins and effectors, iron acquisition, adherence, invasion, and immune system avoidance. Although many paradigms exist that describe mechanisms of pathogenesis, the contribution of microbial metabolism to bacterial virulence during an infection is less understood. Much work has been done studying E. coli as model organism for characterizing individual central metabolism pathways and enzymes [10], [27], [32][38]. We have shown here that central metabolism studies in E. coli can be extended to investigate the contribution of central pathways to bacterial pathogenesis using a virulent uropathogenic E. coli strain and a well-established animal model of UTI. It is known that commensal E. coli require the Entner-Doudoroff pathway and glycolysis for colonization in vivo; while the TCA cycle, pentose phosphate pathway, and gluconeogenesis are dispensable in the intestine [8]. In contrast, we have shown that during E. coli infection of the urinary tract, the pathways required for commensal colonization are dispensable while the TCA cycle and gluconeogenesis are necessary for UPEC fitness in vivo. Adaptation to distinct host environments has been previously shown to involve shared traits between commensal and pathogenic strains [39],[40]. Because commensal E. coli are an important natural component of the intestine one concern faced when developing antimicrobials that target pathogenic strains is how to avoid eradicating commensal bacteria. Thus, these findings highlight important differences between commensal and pathogenic E. coli that could be exploited for the development of antimicrobials that target these pathways in this pathogen during infections that may not affect commensal strains. Interestingly, in addition to UPEC, gluconeogenesis is required for virulence in microbes that represent an array of pathogenic lifestyles, from intracellular bacteria and parasites [41],[42], plant-pathogenic [43], and intestinal pathogens [16]; suggesting that anaplerosis may be a common mechanism of microbial pathogenesis.

This study comprehensively examines the role of pathogenic E. coli central metabolism in a disease model and provides insight not only into UPEC metabolism in vivo but also information regarding the nutrients available to support the growth of E. coli within the urinary tract. The proteomics experiments did reveal that UPEC growing in human urine induces expression of multiple isoforms of both dipeptide- and oligopeptide-binding proteins, both of which were found to be required for UPEC to effectively colonize the urinary tract. This indicates that these bacteria actively import short peptides in urine and this function may indicate that peptides are an important carbon source in vivo. Consistent with this, only bacteria with defects in peptide transport, gluconeogenesis, or the TCA cycle demonstrated a significant reduction in fitness in vivo in both the bladder and kidneys. These findings suggest a model that describes the biochemistry of E. coli during UTI. For optimal growth during infection, short peptides are taken up by UPEC and degraded into amino acids that are catabolized and used in a series of anaplerotic reactions that replenish TCA cycle intermediates and generate gluconeogenesis substrates (Fig. 7).

Figure 7
UPEC acquires amino acids and requires gluconeogenesis and the TCA cycle for fitness in vivo.

Certain glycolytic steps are irreversible and the reverse gluconeogenic reaction is performed by an enzyme specific for gluconeogenesis. Carbon flux through glycolysis and gluconeogenesis must be carefully controlled by the cell to avoid a futile cycle of carbon metabolism [44]. Allosteric regulation of enzymes that catalyze irreversible reactions in these pathways and catabolite repression are mechanisms used to avoid the futile cycle [45],[46]. A gluconeogenic-specific enzyme subject to allosteric regulation is phophoenolpyruvate carboxykinase that converts oxaloacetate to phosphoenolpyruvate [47]. Deletion of the gene pckA that encodes this enzyme resulted in a significant reduction in UPEC fitness in vivo. Because bacteria prevent glycolysis and gluconeogenesis from occurring simultaneously and deletion of pckA reduced fitness in vivo, we reason that carbon flux through gluconeogenesis during UPEC infection may be an important indication of amino acid catabolism in vivo.

It is not surprising that, in addition to gluconeogenesis, the TCA cycle is also required for UPEC fitness in vivo. These two pathways are connected and collectively described as “filling in” or anaplerotic reactions. The TCA cycle is necessary to provide substrates for gluconeogenesis when cells use amino acids as a carbon source. Gluconeogenic amino acids can be degraded to oxaloacetate or to pyruvate that can be converted to acetyl-CoA and enter the TCA cycle [47]. Oxaloacetate, a TCA cycle intermediate, is converted to phophoenolpyruvate during gluconeogenesis by PckA as described above. A mutation in the TCA cycle enzyme succinate dehydrogenase, sdhB, results in a UPEC strain that has reduced fitness in vivo. This finding suggests that UPEC are growing aerobically in the urinary tract because succinate dehydrogenase is replaced by fumarate reductase during anaerobic growth and therefore, future work could confirm if the reductive TCA cycle is not operating during UPEC infection. The requirement for peptide import and the TCA cycle for UPEC fitness during infection is consistent with the hypothesis that acetyl-CoA production from the degradation of amino acids could be occurring in vivo as has been shown by another group [48].

Interestingly, with the exception of peptide-transport proteins, up-regulation of protein expression in urine ex vivo did not correlate with functional importance in vivo. This could be due to the fact that many central metabolism genes are constitutively expressed and that human urine only partially mimics the complex lifestyle of UPEC during UTI [49]. The absence of host cells and the immune response during growth in urine ex vivo could in part account for this discrepancy. It also remains possible that mutants that lack growth defects in urine but demonstrate reduced fitness in vivo could represent genes or metabolic pathways that are required for intracellular phases of growth during cystitis [50].

Despite these disadvantages, up-regulation of both DppA and OppA expression was seen in urine and loss of either dppA or oppA was found to negatively impact UPEC colonization in vivo. Induction of dppA has been reported in a hypervirulent UPEC strain that has a lacks a functional D-serine deaminase gene (dsdA) [51]. Deletion of dppA in this mutant strain resulted in a loss of the hypervirulent phenotype in vivo and significantly reduced its ability to colonize the urinary tract in competition with wild-type [51]. Surprisingly, in contrast to our findings, this group found that mutation of dppA alone had no effect on UPEC fitness in vivo [51]. Due to lack of complementation, it is unclear from that work why loss of dppA dramatically attenuated a hypervirulent strain but had no effect on wild-type. Despite this inconsistency in that work, the importance of peptide transport for UPEC fitness in vivo is supported by the findings that loss of either dppA or oppA significantly reduced colonization of the urinary tract and that the reduced bacterial colonization in the ΔdppA strain can be restored to wild-type levels by complementing the mutant with a wild-type dppA gene.

In summary, defects in the both branches of the pentose phosphate pathway, the Entner-Doudoroff pathway, and glycolysis had limited or no impact on UPEC fitness in vivo. On the other hand, the TCA cycle- and gluconeogenesis-defective strains demonstrate significant fitness reductions during UTI. The utilization of short peptides and amino acids as a carbon source during bacterial infection of the urinary tract is supported by the observation that UPEC mutants defective in peptide import have reduced fitness in vivo while auxotrophic strains do not. Together, these findings provide compelling evidence to support the notion that catabolism of amino acids to form TCA cycle intermediates and gluconeogenic substrates is important for the ability of UPEC to infect the urinary tract efficiently. This shows that anaplerotic and central metabolism pathways are required for UPEC fitness in vivo and suggest microbial metabolism should be considered important for bacterial pathogenesis.

Materials and Methods

Bacteria and growth conditions

Strains were derived from E. coli strain CFT073, a prototypic UPEC strain isolated from the blood and urine of a patient with acute pyelonephritis [52]; its genome has been sequenced and fully annotated [53]. Isolated colonies were used to inoculate overnight Luria-Bertani (LB) cultures. Bacteria from overnight cultures were collected by centrifugation, washed with sterile PBS, and 106 CFU were used to inoculate pre-warmed LB or human urine. To mimic iron-limitation in urine, LB containing 10 mM deferoxamine mesylate (Sigma) was used as a growth medium for comparative proteomics. For human urine cultures, mid-stream urine was collected into sterile sample containers from 8–10 male and female donors, pooled, and sterilized by vacuum filtration through a 0.22 µm pore filter. MOPS defined medium containing 0.2% glucose [54] with and without 10 mM L-arginine, L-serine, glycine, aspartatic acid, or D-serine (Sigma) was also used to test growth of mutant strains. Growth curves were established in triplicate using a Bioscreen bioanalyzer in 0.4 ml volumes; OD600 was recorded every 15 min. All cultures were incubated at 37°C; LB overnight and MOPS cultures were incubated with aeration; urine cultures were incubated statically. For preparation of proteins, UPEC isolate CFT073 was grown statically to exponential phase (OD600 = 0.25) in pre-warmed LB or human urine at 37°C in 5×100 ml cultures for each growth medium.

Preparation of cytoplasmic proteins

Bacteria were harvested from 500 ml of culture by centrifugation (10,000× g, 30 min, 4°C) and lysed in a French pressure cell at 20,000 psi. Harvested cells were washed and resuspended in 10 ml of 10 mM HEPES, pH 7.0 containing 100 U of Benzonase (Sigma). Following two passes through the chilled pressure cell, lysates were centrifuged (7500× g, 10 min, 4°C) to remove unbroken cells and supernatants were ultracentrifuged (120,000× g, 1 h, 4°C) to remove membranes and insoluble material. Soluble proteins were quantified using the 2D Quant Kit (GE Healthcare) following the manufacturer's protocol and either used immediately in DIGE-labeling procedures or stored at −80°C.

2D-DIGE and MS/MS

For fluorescence difference in gel electrophoresis (2D-DIGE) [55], bacterial proteins were minimally labeled with cyanine-derived fluors (CyDyes) containing an NHS ester-reactive group as recommended by the manufacturer (GE Healthcare). To determine quantitative differences within the UPEC soluble proteome during growth in human urine, cytoplasmic proteins prepared from human urine cultures were labeled with Cy3, from LB broth with Cy5, and a pooled internal standard representing equal amounts of both urine and LB preparations with Cy2 as described previously [23]. Briefly, 50 µg of protein was incubated with 400 pmol CyDye for 30 min and the reaction was stopped by added 10 mM lysine. Following labeling, samples labeled with each CyDye were pooled (150 µg total protein), mixed with an equal volume of 2× DIGE sample buffer; 7 M urea, 2 M thiourea, 10 mM tributylphosphine (TBP) (Sigma), 2× biolytes 3–10 (Bio-Rad), 2% ASB-14 and incubated on ice for 10 min. For rehydration, samples were brought to 0.35 ml with 1× DIGE rehydration buffer (7 M urea, 2 M thiourea, 5 mM TBP, 1× biolytes 3–10, 1% ASB-14) and used to passively rehydrate pH 4–7 IPG strips (Bio-Rad) overnight at room temperature. Rehydrated IPG strips were equilibrated and subjected to isoelectric focusing for 50,000 V·h and second dimension SDS-PAGE on 10% gels within low fluorescence glass plates (Jule Biotechnologies, Inc.) and were run at a constant current of 55 mA at 4°C for 4 hr. Following SDS-PAGE, image acquisition and pixel intensity was obtained using a Typhoon scanner (GE Healthcare) and differential in-gel analysis and biological analysis of variance were performed using the DeCyder 6.5 software suite (GE Healthcare). Using this software, the normalized spot volume ratios from Cy3 or Cy5 labeled spots were quantified relative to the Cy2-labeled internal standard from the same gel. The Cy2-labeled standard was then used to standardize and compare normalized volume ratios between the Cy3 and Cy5 labeled proteins between gels representing three independent experiments to generate statistical confidence for abundance changes using student's t-test and ANOVA. To identify the proteins, 500 µg of cytoplasmic proteins were focused as described above and spots of interest were excised from a colloidal Coomassie-stained 2D SDS-PAGE gel and subjected to enzymatic digestion with trypsin. Mass spectra were acquired on an Applied Biosystems 4700 Proteomics Analyzer (TOF/TOF). MS spectra were acquired from 800–3500 Da and the eight most intense peaks in each MS spectrum were selected for MS/MS analysis. Peptide identifications were obtained using GPS Explorer (v3.0, Applied Biosystems), which utilizes the MASCOT search engine. Each MS/MS spectrum was searched against NCBInr. Tryptic digestion and tandem mass spectrometry were performed at the University of Michigan Proteome Consortium.

Construction of UPEC metabolism mutants

Deletion mutants were generated using the lambda red recombinase system [56]. Primers homologous to sequences within the 5′ and 3′ ends of the target genes were designed and used to replace target genes with a nonpolar kanamycin resistance cassette derived from the template plasmid pKD4 [56]. Kanamycin (25 µg/ml) was used for selection of all mutant strains. Gene deletions begin with the start codon and end with the stop codon for each gene. To determine whether the kanamycin resistance cassette recombined within the target gene site, primers that flank the target gene sequence were designed and used for PCR. After amplification, each PCR product was compared to wild-type PCR product and in cases where size-differences are negligible; PCR products were digested with the restriction enzyme EagI (New England Biolabs). Both the PCR products and restriction digests were visualized on a 0.8% agarose gel stained with ethidium bromide. For in vivo complementation, the dppA and pckA genes were amplified from CFT073 genomic DNA using Easy-A high-fidelity polymerase (Stratagene) and independently cloned into pGEN-MCS [28],[57] using appropriate restriction enzymes. The sequences of pGEN-dppA and pGEN-pckA were verified by DNA sequence analysis prior to electroporation into CFT073 ΔdppA or ΔpckA mutant strains.

Experimental UTI

Six-to eight-week-old female CBA/J mice (20 to 22 g; Jackson Laboratories) were anesthetized with ketamine/xylazine and inoculated transurethrally over a 30 sec period with a 50 µl bacterial suspension per mouse using a sterile polyethylene catheter (I.D. 0.28 mm×O.D. 0.61 mm) connected to an infusion pump (Harvard Apparatus). To measure relative fitness, overnight LB cultures for CFT073 and the mutant strain were collected by centrifugation and resuspended in sterile PBS, mixed 1[ratio]1 and adjusted to deliver 2×108 CFU per mouse. Dilutions of each inoculum were spiral plated onto LB with and without kanamycin using an Autoplate 4000 (Spiral Biotech) to determine the input CFU/mL. After 48 hpi, mice were sacrificed by overdose with isoflurane and the bladder and kidneys were aseptically removed, weighed, and homogenized in sterile culture tubes containing 3 ml of PBS using an OMNI mechanical homogenizer (OMNI International). Appropriate dilutions of the homogenized tissue were then spiral plated onto duplicate LB plates with and without kanamycin to determine the output CFU/g of tissue. Plate counts obtained on kanamycin were subtracted from those on plates lacking antibiotic to determine the number of wild-type bacteria. Competitive indices were calculated by dividing the ratio of wild-type to mutant at 48 hpi by the ratio of wild-type to mutant input CFU/mL. Groups of 5 mice per co-challenge were used to determine defects in fitness, when a defect was apparent the co-challenge was repeated two more times with groups of 5 mice. Statistically significant differences in colonization (P-value<0.05) were determined using a two-tailed Wilcoxon matched pairs test. All animal protocols were approved by the University Committee on Use and Care of Animals at the University of Michigan Medical School.

Acknowledgments

The authors would like to thank Daniel Reiss for assisting with the proteomics experiments.

Footnotes

The authors have declared that no competing interests exist.

Funding was provided by Public Health Service grants AI43363 and AI059722 from the National Institutes of Health. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

1. Goering RV, Shawar RM, Scangarella NE, O'Hara FP, Amrine-Madsen H, et al. Molecular epidemiology of methicillin-resistant and methicillin-susceptible Staphylococcus aureus isolates from global clinical trials. J Clin Microbiol. 2008;46:2842–2847. [PMC free article] [PubMed]
2. Dorman SE, Chaisson RE. From magic bullets back to the magic mountain: the rise of extensively drug-resistant tuberculosis. Nat Med. 2007;13:295–298. [PubMed]
3. Pitout JD, Laupland KB. Extended-spectrum beta-lactamase-producing Enterobacteriaceae: an emerging public-health concern. Lancet Infect Dis. 2008;8:159–166. [PubMed]
4. Levy SB, Marshall B. Antibacterial resistance worldwide: causes, challenges and responses. Nat Med. 2004;10:S122–S129. [PubMed]
5. Foxman B, Barlow R, D'Arcy H, Gillespie B, Sobel JD. Urinary tract infection: self-reported incidence and associated costs. Ann Epidemiol. 2000;10:509–515. [PubMed]
6. Litwin MS, Saigal CS, Yano EM, Avila C, Geschwind SA, et al. Urologic diseases in America Project: analytical methods and principal findings. J Urol. 2005;173:933–937. [PubMed]
7. Autieri SM, Lins JJ, Leatham MP, Laux DC, Conway T, et al. L-fucose stimulates utilization of D-ribose by Escherichia coli MG1655 ΔfucAO and E. coli Nissle 1917 ΔfucAO mutants in the mouse intestine and in M9 minimal medium. Infect Immun. 2007;75:5465–5475. [PMC free article] [PubMed]
8. Chang DE, Smalley DJ, Tucker DL, Leatham MP, Norris WE, et al. Carbon nutrition of Escherichia coli in the mouse intestine. Proc Natl Acad Sci U S A. 2004;101:7427–7432. [PMC free article] [PubMed]
9. Fabich AJ, Jones SA, Chowdhury FZ, Cernosek A, Anderson A, et al. Comparison of carbon nutrition for pathogenic and commensal Escherichia coli strains in the mouse intestine. Infect Immun. 2008;76:1143–1152. [PMC free article] [PubMed]
10. Peekhaus N, Conway T. What's for dinner? Entner-Doudoroff metabolism in Escherichia coli. J Bacteriol. 1998;180:3495–3502. [PMC free article] [PubMed]
11. Miranda RL, Conway T, Leatham MP, Chang DE, Norris WE, et al. Glycolytic and gluconeogenic growth of Escherichia coli O157[ratio]H7 (EDL933) and E. coli K-12 (MG1655) in the mouse intestine. Infect Immun. 2004;72:1666–1676. [PMC free article] [PubMed]
12. Bonafonte MA, Solano C, Sesma B, Alvarez M, Montuenga L, et al. The relationship between glycogen synthesis, biofilm formation and virulence in Salmonella enteritidis. FEMS Microbiol Lett. 2000;191:31–36. [PubMed]
13. Jones SA, Jorgensen M, Chowdhury FZ, Rodgers R, Hartline J, et al. Glycogen and maltose utilization by Escherichia coli O157[ratio]H7 in the mouse intestine. Infect Immun. 2008;76:2531–2540. [PMC free article] [PubMed]
14. McMeechan A, Lovell MA, Cogan TA, Marston KL, Humphrey TJ, et al. Glycogen production by different Salmonella enterica serotypes: contribution of functional glgC to virulence, intestinal colonization and environmental survival. Microbiology. 2005;151:3969–3977. [PubMed]
15. Mercado-Lubo R, Gauger EJ, Leatham MP, Conway T, Cohen PS. A Salmonella enterica serovar typhimurium succinate dehydrogenase/fumarate reductase double mutant is avirulent and immunogenic in BALB/c mice. Infect Immun. 2008;76:1128–1134. [PMC free article] [PubMed]
16. Tchawa Yimga M, Leatham MP, Allen JH, Laux DC, Conway T, et al. Role of gluconeogenesis and the tricarboxylic acid cycle in the virulence of Salmonella enterica serovar Typhimurium in BALB/c mice. Infect Immun. 2006;74:1130–1140. [PMC free article] [PubMed]
17. Brooks T, Keevil CW. A simple artificial urine for the growth of urinary pathogens. Lett Appl Microbiol. 1997;24:203–206. [PubMed]
18. Snyder JA, Haugen BJ, Buckles EL, Lockatell CV, Johnson DE, et al. Transcriptome of uropathogenic Escherichia coli during urinary tract infection. Infect Immun. 2004;72:6373–6381. [PMC free article] [PubMed]
19. Anfora AT, Haugen BJ, Roesch P, Redford P, Welch RA. Roles of serine accumulation and catabolism in the colonization of the murine urinary tract by Escherichia coli CFT073. Infect Immun. 2007;75:5298–5304. [PMC free article] [PubMed]
20. Roesch PL, Redford P, Batchelet S, Moritz RL, Pellett S, et al. Uropathogenic Escherichia coli use d-serine deaminase to modulate infection of the murine urinary tract. Mol Microbiol. 2003;49:55–67. [PubMed]
21. Bahrani-Mougeot FK, Buckles EL, Lockatell CV, Hebel JR, Johnson DE, et al. Type 1 fimbriae and extracellular polysaccharides are preeminent uropathogenic Escherichia coli virulence determinants in the murine urinary tract. Mol Microbiol. 2002;45:1079–1093. [PubMed]
22. Russo TA, Jodush ST, Brown JJ, Johnson JR. Identification of two previously unrecognized genes (guaA and argC) important for uropathogenesis. Mol Microbiol. 1996;22:217–229. [PubMed]
23. Alteri CJ, Mobley HL. Quantitative profile of the uropathogenic Escherichia coli outer membrane proteome during growth in human urine. Infect Immun. 2007;75:2679–2688. [PMC free article] [PubMed]
24. Russo TA, Carlino UB, Mong A, Jodush ST. Identification of genes in an extraintestinal isolate of Escherichia coli with increased expression after exposure to human urine. Infect Immun. 1999;67:5306–5314. [PMC free article] [PubMed]
25. Letoffe S, Delepelaire P, Wandersman C. The housekeeping dipeptide permease is the Escherichia coli heme transporter and functions with two optional peptide binding proteins. Proc Natl Acad Sci U S A. 2006;103:12891–12896. [PMC free article] [PubMed]
26. Sprenger GA. Genetics of pentose-phosphate pathway enzymes of Escherichia coli K-12. Arch Microbiol. 1995;164:324–330. [PubMed]
27. Fraenkel DG, Vinopal RT. Carbohydrate Metabolism in Bacteria. Annu Rev Microbiol. 1973;27:69–100.
28. Lane MC, Alteri CJ, Smith SN, Mobley HL. Expression of flagella is coincident with uropathogenic Escherichia coli ascension to the upper urinary tract. Proc Natl Acad Sci U S A. 2007;104:16669–16674. [PMC free article] [PubMed]
29. Irani M, Maitra PK. Isolation and characterization of Escherichia coli mutants defective in enzymes of glycolysis. Biochem Biophys Res Commun. 1974;56:127–133. [PubMed]
30. Chao YP, Patnaik R, Roof WD, Young RF, Liao JC. Control of gluconeogenic growth by pps and pck in Escherichia coli. J Bacteriol. 1993;175:6939–6944. [PMC free article] [PubMed]
31. Hanson RS, Cox DP. Effect of different nutritional conditions on the synthesis of tricarboxylic acid cycle enzymes. J Bacteriol. 1967;93:1777–1787. [PMC free article] [PubMed]
32. Babul J, Clifton D, Kretschmer M, Fraenkel DG. Glucose metabolism in Escherichia coli and the effect of increased amount of aldolase. Biochemistry. 1993;32:4685–4692. [PubMed]
33. Fraenkel DG. Mutants in glucose metabolism. Annu Rev Biochem. 1986;55:317–337. [PubMed]
34. Fraenkel DG, Levisohn SR. Glucose and gluconate metabolism in an Escherichia coli mutant lacking phosphoglucose isomerase. J Bacteriol. 1967;93:1571–1578. [PMC free article] [PubMed]
35. Josephson BL, Fraenkel DG. Sugar metabolism in transketolase mutants of Escherichia coli. J Bacteriol. 1974;118:1082–1089. [PMC free article] [PubMed]
36. Kim J, Copley SD. Why metabolic enzymes are essential or nonessential for growth of Escherichia coli K12 on glucose. Biochemistry. 2007;46:12501–12511. [PubMed]
37. Vinopal RT, Hillman JD, Schulman H, Reznikoff WS, Fraenkel DG. New phosphoglucose isomerase mutants of Escherichia coli. J Bacteriol. 1975;122:1172–1174. [PMC free article] [PubMed]
38. Zablotny R, Fraenkel DG. Glucose and gluconate metabolism in a mutant of Escherichia coli lacking gluconate-6-phosphate dehydrase. J Bacteriol. 1967;93:1579–1581. [PMC free article] [PubMed]
39. Sokurenko EV, Chesnokova V, Dykhuizen DE, Ofek I, Wu XR, et al. Pathogenic adaptation of Escherichia coli by natural variation of the FimH adhesin. Proc Natl Acad Sci U S A. 1998;95:8922–8926. [PMC free article] [PubMed]
40. Sokurenko EV, Feldgarden M, Trintchina E, Weissman SJ, Avagyan S, et al. Selection footprint in the FimH adhesin shows pathoadaptive niche differentiation in Escherichia coli. Mol Biol Evol. 2004;21:1373–1383. [PubMed]
41. Liu K, Yu J, Russell DG. pckA-deficient Mycobacterium bovis BCG shows attenuated virulence in mice and in macrophages. Microbiology. 2003;149:1829–1835. [PubMed]
42. Naderer T, Ellis MA, Sernee MF, De Souza DP, Curtis J, et al. Virulence of Leishmania major in macrophages and mice requires the gluconeogenic enzyme fructose-1,6-bisphosphatase. Proc Natl Acad Sci U S A. 2006;103:5502–5507. [PMC free article] [PubMed]
43. Liu P, Wood D, Nester EW. Phosphoenolpyruvate carboxykinase is an acid-induced, chromosomally encoded virulence factor in Agrobacterium tumefaciens. J Bacteriol. 2005;187:6039–6045. [PMC free article] [PubMed]
44. Chambost JP, Fraenkel DG. The use of 6-labeled glucose to assess futile cycling in Escherichia coli. J Biol Chem. 1980;255:2867–2869. [PubMed]
45. Garfinkel L, Kohn MC, Garfinkel D. Computer simulation of the fructose bisphosphatase/phosphofructokinase couple in rat liver. Eur J Biochem. 1979;96:183–192. [PubMed]
46. Koerner TA, Jr, Voll RJ, Younathan ES. A proposed model for the regulation of phosphofructokinase and fructose 1,6-bisphosphatase based on their reciprocal anomeric specificities. FEBS Lett. 1977;84:207–213. [PubMed]
47. Gottschalk G. Bacterial Metabolism. New York: Springer-Verlag; 1986.
48. Anfora AT, Halladin DK, Haugen BJ, Welch RA. Uropathogenic Escherichia coli CFT073 is adapted to acetatogenic growth but does not require acetate during murine urinary tract infection. Infect Immun. 2008;76:5760–5767. [PMC free article] [PubMed]
49. Kau AL, Hunstad DA, Hultgren SJ. Interaction of uropathogenic Escherichia coli with host uroepithelium. Curr Opin Microbiol. 2005;8:54–59. [PubMed]
50. Reigstad CS, Hultgren SJ, Gordon JI. Functional genomic studies of uropathogenic Escherichia coli and host urothelial cells when intracellular bacterial communities are assembled. J Biol Chem. 2007;282:21259–21267. [PubMed]
51. Haugen BJ, Pellett S, Redford P, Hamilton HL, Roesch PL, et al. In vivo gene expression analysis identifies genes required for enhanced colonization of the mouse urinary tract by uropathogenic Escherichia coli strain CFT073 dsdA. Infect Immun. 2007;75:278–289. [PMC free article] [PubMed]
52. Mobley HL, Green DM, Trifillis AL, Johnson DE, Chippendale GR, et al. Pyelonephritogenic Escherichia coli and killing of cultured human renal proximal tubular epithelial cells: role of hemolysin in some strains. Infect Immun. 1990;58:1281–1289. [PMC free article] [PubMed]
53. Welch RA, Burland V, Plunkett G, 3rd, Redford P, Roesch P, et al. Extensive mosaic structure revealed by the complete genome sequence of uropathogenic Escherichia coli. Proc Natl Acad Sci U S A. 2002;99:17020–17024. [PMC free article] [PubMed]
54. Neidhardt FC, Bloch PL, Smith DF. Culture medium for enterobacteria. J Bacteriol. 1974;119:736–747. [PMC free article] [PubMed]
55. Unlu M, Morgan ME, Minden JS. Difference gel electrophoresis: a single gel method for detecting changes in protein extracts. Electrophoresis. 1997;18:2071–2077. [PubMed]
56. Datsenko KA, Wanner BL. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci U S A. 2000;97:6640–6645. [PMC free article] [PubMed]
57. Galen JE, Nair J, Wang JY, Wasserman SS, Tanner MK, et al. Optimization of plasmid maintenance in the attenuated live vector vaccine strain Salmonella typhi CVD 908-htrA. Infect Immun. 1999;67:6424–6433. [PMC free article] [PubMed]

Articles from PLoS Pathogens are provided here courtesy of Public Library of Science
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

  • MedGen
    MedGen
    Related information in MedGen
  • PubMed
    PubMed
    PubMed citations for these articles

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...