Our studies demonstrate that growth in urine may be associated with endogenous oxidative stress. It is well known that urine supports bacterial growth. Several studies have shown that UPEC strains grow well in human urine, whereas faecal isolates tended to grow more poorly [19
]. Other studies have also reported that ABU isolates grow faster than UPEC strains [11
]. However, Alteri and Mobley have recently shown that growth in urine is not restricted to UPEC bacteria or ABU strains. Commensal and enteropathogen E. coli
strains produced growth curves indistinguishable from those of UPEC [42
]. In agreement with their work, we found that the ability to grow of twenty-one E. coli
belonging to different pathovars and phylogenetic groups was nearly similar in pooled human urine. The slow growth of a few isolates could reflect the RpoS polymorphism in E. coli
Cultures were grown in microaerophilic conditions (see Methods), where oxygen levels are similar to those in the human urinary tract [18
]. This reduction in oxygen content leads to a redistribution of metabolic fluxes between fermentation and respiration [46
]. Such a shift may decrease the respiratory chain-mediated generation of ROS. Moreover, in our culture conditions, autoxidizable enzymes such as L-aspartate oxidase and fumarate reductase should not contribute significantly to the formation of H2
However, metabolic reactions that generate nearly two-thirds of H2
are not yet identified [47
]. Therefore, we can expect that changes in metabolic fluxes generate different ROS levels. Analysis of metabolic capabilities in a collection of 153 E. coli
natural isolates [48
] and of gene expression in strains ABU 83972 and CFT073 grown exponentially in urine [49
] revealed significant differences in their metabolic capacities. These metabolic changes could therefore generate different ROS levels in our isolates.
Urine is a complex growth medium and E. coli
must adapt to stress imposed by this tough environment. The high osmolality, high urea concentration, low pH and the limitation of certain components could provoke an oxidative response. To protect from these highly reactive intermediates, cells possess a defense system consisting of both enzymatic and non- enzymatic antioxidants that scavenge them. Nevertheless, under several situations, the rate of generation of ROS exceeds that of their removal and oxidative stress occurs. The levels of damage products accumulated (estimated as TBARS concentration) mirror the intensity of oxidative stress. Our results demonstrate that E. coli
strains can respond very differently to stress imposed by urine. TBARS measurements revealed that many E. coli
are exposed to ROS during exponential growth in urine. Surprisingly, this is the case of ABU strain 83972 that is very well adapted to growth in the urinary tract [11
]. In contrast, two other ABU strains 38 and 62, as UTI89, Sakai and MG1655 showed a lower oxidative damage to lipid. No clear correlation between ROS level and the phylogroups or pathogenic group was apparent.
ABU isolates form a heterogenous group. Individual ABU strains display many differences between them in their genome contents and in virulence-associated genes such as LPS, microcin, aerobactin, and mobility [11
]. Interestingly, our study shows that the two ABU strains (38 and 62) belonging to the group B1, differ from other by the low ROS production in urine.
The commensal-like ABU 83972 strain and the pathogenic strain CFT073 are very closely related and belong to the same B2 subgroup II [25
], or to the same sequence type 73 clonal group [4
]. These both strains are genetically very similar and differ only by a few hundred genes [50
]. However, strain ABU 83972 is able to outcompete CFT073 strain in urine [51
]. The results presented herein indicate that both strains undergo an oxidative stress during the exponential growth. Nonetheless, ABU 83972 strain displays more active antioxidant defenses which led to a significant decrease in ROS level in stationary phase. Our results agree with the gene expression profiling in strains ABU 83972 and CFT073 in urine, which showed that sodA
, encoding superoxide dismutase and ahpC
, encoding hydroperoxide reductase are significantly up-regulated [49
]. Interestingly the highest expression values were obtained in ABU 83972 strain [49
]. To further explore the oxidative response, other studies will be performed to examine the contribution of each factor involved in this response and the importance of metabolic changes in these isolates.
The UPEC strains CFT073 (urosepsis/pyelonephritis isolate), 536 (pyelonephritis, B2 subgroup III) and UTI89 (cystitis, B2 subgroup IX) [25
] are very well adapted for growth in the human urinary tract and present similar antioxidant defense systems. However, a clear distinction can be drawn between them. Strains CFT073 and 536 behave similarly with respect to ROS formation in exponential phase in contrast to UTI89 (p
0.016). The metabolic fluxes could be distributed differently in UTI89, which may decrease the endogenous production of ROS.
The more efficient antioxidant metabolism related to greater exposure to endogenous oxidative stress may be responsible for the difference in lifestyle between ABU 83972 and CFT073 strains. ABU 83972 strain exploits urine more efficiently than UPEC strains [11
]. Previous study has shown that a more active antioxidant defense system increases the capacity to colonize the bladder [53
]. Thus, a high level of antioxidant defenses associated to fast growth in the urine (this work), low abundance of fimbriae, and possible biofilm formation [54
] could explain why ABU83972 strain is able to establish a long-term bacteriuria. Additionally, ROS are implicated in DNA mutagenesis which may be adaptive as reported in biofilm for antibiotic-resistance [55
], or more generally, during starvation [56
]. The high levels of ROS in ABU strain 83972 may explain the genetic alterations described [27