«(Über die Bedeutung der bakteriellen Genomplastizität für die Adaptation und Evolution asymptomatischer Bakteriurie (ABU) Escherichia coli ...»
Bacterial growth in Luria broth resulted in similar results relative to growth in urine where most of the re-isolates grew worse with respect to growth rate and optical density reached compared to the parent strain (Table 11). Re-isolates of patient SR, however, grew as well as parent strain 83972 in LB medium (data not shown). However, the differences were most prominent for re-isolates of patients KA, IJ and POS.
These data may indicate a possible correlation between the growth characteristics of reisolates and the immune response of their specific host. In general, re-isolates from patients with a stronger host response grew more slowly than re-isolates from patients with weaker host response. However, bacterial growth characteristics as well as host response are very complex phenomena and differ from patient to patient.
Fig. 27: Growth characteristics of in vivo re-isolates of strain 83972 and their parent strain in pooled human urine.
Competitiveness To test whether the re-isolates differ with respect to competitiveness in urine from their parent strain, E. coli strain 83872 was tagged with a chloramphenicol resistance cassette. Using the λ Red-mediated integration of linear DNA fragments into the bacterial chromosome (Datsenko and Wanner, 2000), the cat cassette was integrated into the bacteriophage λ chromosomal attachment site of strain 83972.
For growth competition experiments, identical bacterial numbers of strain 83972cat and one re-isolate were mixed and grown for 72 h in pooled human urine. At different time points (6, 24, 48 and 72 hours), the ratio of the parent strain 83972cat and the re-isolate was determined by counting colony forming units (CFUs) on LB agar plates supplemented with chloramphenicol and LB plates, respectively. The results of the growth competition experiments with the different re-isolates are shown in Fig. 28.
Fig. 28: Competitiveness in urine of in vivo re-isolates of E. coli 83972 against their parent strain derivative 83972cat. All experiments were performed in triplicates. The last graph shows the control experiment where the parent strain and its chloramphenicol-resistant variant 83972cat have been cocultured.
In a control experiment, the parent strain 83972 was co-cultured with its chloramphenicolresistant derivative 83972cat which did not show any difference in competitiveness thus confirming that the introduction of the cat cassette had no negative effect on the growth rate.
However, the re-isolates co-cultured in urine with strain 83972cat exhibited differences in their competitiveness. The most striking difference was observed for the re-isolates KA22 and KA25 which represented only 25 % or less of the total culture after 72 h of growth. Similarly, the fraction of re-isolates IJ15 and SN16 already decreased significantly after 6 h of cocultivation with strain 83972cat. With exception of strain SN16, this could be correlated with the reduced growth of these strains relative to their parent strain (Fig. 27). Most of the reisolates from patient POS as well as strain CK9 were also outcompeted to a different extend by the strain 83972cat. Generally, re-isolates which were shown to be less competitive were also characterized by slower growth rates in urine relative to the parent strain.
It has been published that increased biofilm formation is characteristic for ABU isolates and that this trait may be important for colonization of the urinary tract (Hancock et al., 2007).
Consequently, re-isolates obtained from the human colonization study were investigated for their biofilm forming ability. For that purpose, bacteria were grown in microtiter plates in pooled human urine, as well as in M63 minimal medium, and the biofilm formation was
compared to that of strain 83972 (section 4.5.4.). In addition, UPEC strain 536 and nonpathogenic laboratory K-12 strain MG1655 were included as positive and negative controls, respectively (Fig. 29).
Fig. 29: Analysis of biofilm formation of in vivo re-isolates of ABU strain 83972 at 37 °C. A) Biofilm formation in pooled human urine; B) Biofilm formation in M63 media. Biofilm formation of the strains tested was normalized to that of parent strain 83972.
Surprisingly, except for strain CK12, all re-isolates formed poorly biofilm in urine when compared to strain 83972. Moreover, the ability to form biofilms of re-isolates SN25, CK6, SR12 and POS18 was as low as or less than that of non-pathogenic K-12 strain MG1655. The ability to form biofilms of the other seven re-isolates (KA25, CK3, SR3, SR6, POS6, POS9, POS12) was comparable to that of UPEC strain 536. In general, most of the tested re-isolates were not able to form as good biofilm as the ancestor strain 83972. In contrast, the biofilm assays in M63 minimal medium showed very distinct results. Except for strains CK6, CK12 and SR12, biofilm formation of the re-isolates was comparable to that of their parent strain
83972. Re-isolate CK6 formed nearly as much biofilm as strain 536, which was 7-fold higher than that of strain 83972.
In general, biofilm formation could be correlated with the swarming ability (Fig. 26). Strains with increased motility formed less biofilm than strains exhibiting low motility. The results of the phenotypic characterization of the in vivo re-isolates of strain 83972 are summarized in Table 10.
Differences in the restriction pattern are indicated by “+”. An increase or decrease of the traits tested relative to ABU strain 83972 is indicated by corresponding arrows; no differences between re-isolate and parent strain are indicated by “=”.
Taken together, these date demonstrate differences in phenotypes of strain 83972 in vivo reisolates. Consequently, growth characteristics were affected and almost all re-isolates did not reach the same final bacterial number and had lower growth rates in the urine. Generally, reisolates that were characterized by slower growth rates in urine, were shown to be less competitive relative to the parent strain. These data also demonstrate that strain 83972 is capable of modulating the swarming ability in response to the growth environment.
Furthermore, most of the tested re-isolates were not able to form as good biofilm as the parent strain 83972 and this phenotype could be correlated with the swarming ability.
5.3.6. Host independent growth of E. coli strain 83972 Bacterial growth is modulated by several environmental factors like competition for nutrients, niche-specific conditions, or host response. In order to identify host factors that affect bacterial growth and adaptation, an in vitro continuous culture system was designed in which cultures of strain 83972 could be propagated for more than 2000 generations without an impact of host factors. Consecutive in vitro re-isolates of strain 83972 were taken once a week and characterized with regard to their pheno- and genotypic properties.
slightly varied, what could be due to sampling errors. Nevertheless, an increase in OD 600 always corresponded to an increase in CFU.
Fig. 30: Growth dynamics of E. coli strain 83972 during continous culture experiments in LB medium + NO A), urine + NO B), LB medium C) and urine D).
At each sampling time point, bacteria were tested for spontaneous occurrence of resistance against streptomycin in order to monitor the occurrence of mutator phenotypes. Only three
resistant isolates were found, however, at different sampling points. Interestingly, all of the resistant clones were obtained from the urine + NO culture, after 1, 2 and 6 weeks of cultivation.
Moreover, biofilm formation of strain 83972 differed under these four growth conditions (Fig.
31). The addition of exogenous nitric oxide significantly decreased biofilm formation without affecting bacterial growth (Fig. 30). Scanning electron microscopy indicated that the drastic differences could be due to extracellular matrix production. Comparing samples from the urine cultures with and without nitric oxide indicated that the biofilm structure without exposure to nitric oxide is more homogenous and less densely packed with crystals and extracellular matrix (Fig. 31 B and D).
Fig. 31: Biofilm formation of ABU strain 83972 in continous cultures in LB medium + NO A), urine + NO B), LB medium C) and urine D). In each section of the figure: left – scanning electron micrograph of the biofilm architecture; right – overview of biofilm formation in the corresponding microfermenter.
Taken together, an in vitro continuous culture system was designed in which strain 83972 was successfully propagated for more than 2000 generations and later on compared to those reisolates obtained from patient colonisation study.
5.3.7. Genomic and phenotypic properties of ABU strain 83972 grown in vitro In order to characterise genotypic and phenotypic properties of the strain 83972 after long term in vitro continuous culture, single in vitro re-isolates were defined. For this purpose, 17 independent colonies from every microfermenter were picked at the last sampling event after more than 2000 generations. Every colony was subsequently grown overnight in the same medium in which it has been propagated before. Bacteria from these cultures which represented an individual 83972 in vitro re-isolate were stored in glycerol stocks at -80 °C for later experiments.
Genetic structure of the in vitro 83972 re-isolates
According to restriction pattern determined by pulsed-field gel electrophoresis following digestion with XbaI or AvrII, the genetic structure of all in vitro re-isolates was not altered and was identical to that of strain 83972 used for initial inoculation of the fermenters (Fig.
32). This discovery is in contrast to the results obtained for the genome structure of in vivo E.
coli 83972 re-isolates where multiple genome rearrangements were described even after less generations of growth in the human bladder (Fig. 25).
Interestingly, even the addition of mutagenic nitric oxide did not result in genomic alterations in E. coli strain 83972. The fact that growth in the human bladder induced genomic alterations detectable by PFGE, but not in vitro culture in pooled human urine, indicates that specific conditions exist within the human urinary tract which are driving forces of bacterial genome plasticity.
Fig. 32: Comparison of the genome structure of in vitro re-isolates of E. coli strain 83972 by PFGE following restriction with XbaI (left column) and AvrII (right column). 17 independent colonies were picked after more than 2000 generations in continous cultures in LB medium + NO A), urine + NO B), LB medium C) and urine D).
Motility As already described before (section 5.3.5.), E. coli strain 83972 is able to modulate the expression of flagella. The comparison of motility of in vitro re-isolates corroborated these results. In general, growth in Luria broth resulted more frequently in the occurrence of isolates with a changed motility relative to the parent strain 83972 than growth in urine.
Interestingly, increased motility was only observed among urine culture isolates. In LB + NO culture, 1/3 of bacterial isolates showed decrease in motility, whereas in LB without NO culture already 2/3 isolates were non-motile. Similarly, in the urine cultures, the addition of NO decreased double the number of non-motile 83972 isolates (Table 12). Interestingly, in LB and urine cultures without addition of NO, decrease in motility could be correlated with increased whole culture population biofilm formation (Fig. 31).
Growth characteristics In addition, the growth characteristics of the in vitro re-isolates were tested in pooled human urine. In general, growth of all tested in vitro re-isolates did not differ markedly from those of their parent strain 83972 (Fig. 33). Nevertheless, most of the re-isolates from urine cultures
exhibited a slightly increased growth rate. Differences in the growth rates of re-isolates from LB cultures were always in range of the standard deviation.
Fig. 33: Growth characteristics of in vitro re-isolates of strain 83972 and their parent strain in pooled human urine at 37 °C. 17 independent re-isolates were selected after more than 2000 generations in continous cultures in LB medium + NO A), urine + NO B), LB medium C) and urine D).
Similarly, the comparison of biofilm formation by the in vitro re-isolates resulted in a very few significant differences when compared to that of parent strain 83972 (Fig. 32). If biofilm formation of the in vitro re-isolates was altered relative to strain 83972, increased biofilm formation was observed mainly for isolates from LB cultures. Noteworthy, this observation is
Fig. 34: Analysis of biofilm formation of in vitro re-isolates of ABU strain 83972 in pooled human urine at 37 °C. 17 independent re-isolates were selected after more than 2000 generations in continous cultures in LB medium + NO A), urine + NO B), LB medium C) and urine D). Biofilm formation of the strains tested was normalized to that of parent strain 83972.