«(Über die Bedeutung der bakteriellen Genomplastizität für die Adaptation und Evolution asymptomatischer Bakteriurie (ABU) Escherichia coli ...»
Another example of an alternative nutritional strategy of strain 83972 during growth in the bladder is the degradation of ribo- and deoxyribonucleosides. This is apparent from the proteome and transcriptome comparison of in vivo re-isolate CK12 and its parent strain
83972. Nucleic acids are highly abundant in the urine when many epithelial cells undergo apoptosis and their DNA is degraded. Also bacterial lysis is a good source of freely available nucleic acids in the urine. In the re-isolate, the gene udp and its product uridine phosphorylase were found to be strongly up-regulated by transcriptome and cytoplasmic proteome comparison, respectively. This enzyme catalyses phosphorylation of uridine followed by the conversion to ribose and uracil. Uridine diphosphate serves as a glycosyl carrier in many important reactions like glucuronidation, where UDP-glucuronic acid (glucuronic acid linked via a glycosidic bond to uridine diphosphate) is an intermediate in the process (King et al., 2000). In the animal body, glucuronic acid is often linked to the xenobiotic metabolism of substances such as drugs, pollutants, bilirubin, androgens, estrogens, mineralocorticoids, glucocorticoids, fatty acid derivatives, retinoids, and bile acids. The substances resulting from glucuronidation are known as glucuronides (or glucuronosides) and are typically much more water-soluble than the non-glucuronic acid-containing substance from which they were
originally synthesised. The human body uses glucuronidation to make a large variety of substances more water-soluble, and in this way, allow for their subsequent elimination from the body upon urination (King et al., 2000). Most likely bacteria might be able to employ enzymes like Udp to use uridine as a carbon source via the pentose phosphate pathway.
One more example how bacteria utilize compounds from urine as a carbon source is D-serine metabolism. D-serine is excreted in human urine at concentrations ranging from 3.0 to 40 µg ml-1 (Brückner et al., 1994). An epidemiological study demonstrated that the dsdA gene encoding for the D-serine deaminase is more frequently present in uropathogenic than in faecal isolates (Roesch et al., 2003). Indeed, the asymptomatic strain 83972 possesses the dsdA gene and, depending on the colonized host, might use it to certain extent. In case of reisolate SR12, transcriptome analysis demonstrated that the genes dsdA and dsdX (encoding for D-serine transport) were significantly up-regulated compared to the parent strain 83972.
Moreover, genes coding for proteins involved in L-serine metabolism were found to be downregulated indicating that D-serine but not L-serine is used by E. coli strain SR12 as a carbon source. Interestingly, D-serine inhibits in higher concentrations growth of E. coli by blocking the L-serine and pantothenate biosynthesis (Cosloy and McFall, 1973). Therefore, upregulation of D-serine catabolism by E. coli strains colonizing the urinary tract might have a dual function: nutrition and detoxification.
Recent findings further underline the importance of D-serine during urinary tract colonisation in addition to the nutritional aspect. In vitro transcriptome analysis of UPEC strain CF073 and its dsdA mutant during murine infection revealed that a set of genes coding for virulence factors including P- and F1C fimbriae as well as alpha-hemolysin were up-regulated in the mutant. Moreover the dsdA mutant was hyperflagellated and outcompetes the wild type strain in the murine urinary tract (Haugen et al., 2007). dsdA expression is phase variable and this switch is linked to the expression of type 1 fimbriae and reciprocal to motility (Anfora et al., 2007). Altogether, these findings underline the importance of serine homeostasis in the bacterial cell during growth in the urinary tract. According to this model, already inactivated virulence-associated genes in ABU strain 83972 would be further down-regulated due to upregulation of dsdA expression. Furthermore, increased D-serine transport into the cell would explain the increased flagella expression in re-isolate SR12.
D-serine deamination results in pyruvate and ammonia (Roesch et al., 2003) and the latter product can be used as a nitrogen source. In line with that, the transcriptome analysis of reisolate SR12 revealed along with the up-regulation of genes required for D-serine catabolism, induction of many genes involved in nitrogen homeostasis (Fig. 51). There are two physiologically independent pathways of ammonium assimilation in E. coli with glutamate and glutamine as primary products (Reitzer, 2003). Glutamate synthesis from α-ketoglutarate and ammonia (via the GDH pathway) is employed under energy-limited (presumably nitrogen-rich) conditions, whereas glutamine synthesis (via the GS-GOGAT pathway) is used under energy-rich conditions and consumes ATP (Helling, 1994). According to transcriptome data, the first pathway is particularly used by strain SR12 since the expression of glutamate dehydrogenase and glutamine synthetase-encoding genes was found to be up-regulated and down-regulated, respectively (Fig. 51). In addition, expression of the two-component system GlnGL, which is involved in response to nitrogen limitation and which positively regulates glutamine synthesis (Reitzer, 2003), was also found to be down-regulated on the transcriptional level. Moreover, transcriptome profiling uncovered glutamine transport to be repressed in strain SR12 relative to parent strain 83972.
Taken together, these data demonstrate that during growth in the urinary tract the E. coli metabolism and catabolism are efficiently adjusted to the specific nutrients supplied. Strain 83972 is able to take up and metabolize a number of sugars, sugar-derivatives and amino acids present in urine. Energy metabolism and nitrogen assimilation seem to be adjusted to the low energy and nitrogen-rich growth conditions in urine.
6.7. Outer membrane protein profile and iron uptake
For bacteria living within other organisms, the outer membrane is a critical barrier that directly interacts with the host components (Cullen et al., 2004). Depending on the bacterial colonization strategy, either many cell surface-associated proteins including adhesins will be expressed or only a minimal protein set that facilitates the acquisition of nutrients and macromolecules from the respective niche. E. coli associated with asymptomatic bacteriuria would be expected to fall into the second group because outer membrane proteins are known
to induce host response and immunogenic reactions (Hagan and Mobley, 2007). To characterise the expression of cell surface-associated structures, ABU strain 83972 was analysed with regard to the outer membrane protein (OMP) profile.
2D gel electrophoresis (2D-GE) is a powerful tool to investigate the composition of protein fractions. Unfortunately, there are some drawbacks like protein solubility and their hydrophobic character. Another critical step is protein isolation (Molloy et al., 2000), as, depending on the cell lysis strategy, most of the loosely surface-associated proteins such as fimbrial components are removed and the remaining fraction consists of mainly integral OMPs. As the determinants coding for typical UPEC adhesins like F1C, P, type 1 fimbriae are non-functional in strain 83972 (Klemm et al., 2006; Roos et al., 2006a), and due to the bacterial cell lysis using the French press, it was not expected to detect these adhesins by means of 2D-GE.
The three in vivo re-isolates (CK12, SR12 and KA25) and their parent strain 83972 exhibited very similar OMP profiles. Neverhtless, certain alterations in individual protein amounts were observed. Of the 18 detected proteins in parent strain 83972, eight proteins (FepA, FecA, FhuA, FhuE, ChuA, IutA, IroN and Iha) are known components of siderophore systems.
Among the proteins which are not involved in iron uptake, OmpA and OmpC the major porins involved in diffusion and influx of nutrients into the periplasm (Garten et al., 1975) were the most abundant ones. This data further corroborate our expectations that the strain 82972 is well adapted to growth in urine as an iron-limited medium. The expression of multiple iron uptake systems is therefore fundamental to efficiently grow in this niche (Andrews et al., 2003; Baumler et al., 1996; Torres and Payne, 1997).
E. coli surface-exposed proteins are frequently anchored in the outer membrane and serve as antigens for the human immune system. It is expected that individual patients differ markedly in their efficiency to express defence mechanisms against bacteria. To analyse the impact of prolonged host-pathogen interaction, the outer membrane protein profile of the in vivo reisolates from individual patients were compared. The most abundant and rather constitutively expressed protein among the tested re-isolates was FepA, which is involved in transport of enterobactin-bound iron across the outer membrane (Sansom, 1999). The second most accumulated protein was IroN. In contrast to FepA, IroN amounts varied from strain to strain.
Hagan and Mobley (2007) reported IroN to be immunogenic in mice. In the same study, the
proteins IutA, ChuA and Iha were mentioned to be antigenic in the murine model. We found out that IutA and Iha were differently expressed in the patients and that they were not expressed at all in re-isolate CK12. This shows that expression of multiple iron uptake systems in strain 83972 varies depending on the colonized host. Referring to the work of Hagan and Mobley (2007), many surface-associated proteins are immunogenic. Therefore, we propose that the quality of the host immune response plays an important role and directs the optimal expression of individual iron uptake systems under these conditions.
6.8. Host defence-driven bacterial gene expression
Changes in bacterial gene expression upon contact with the human host are intensely studied during the last years. This co-existence does also affect the host physiology. Even ABU isolates do induce host immune response, which involves chemokine production and neutrophile influx at the site of infection (Haraoka et al., 1999; Samuelsson et al., 2004).
Neutrophiles are known to produce and release significant amounts of nitric oxide (NO), which is, however, also produced by other mammalian cells (Bogdan, 2001). Therefore, NO functions not only as an antimicrobial agent but also has a pleiotropic effect in the human body (i.e. signalling function) (Bogdan, 2001). It is interesting to understand how eukaryotic cells switch off these signals or protect themselves from NO and NO-related molecules produced for defence purposes. Recently, it has been shown that in eukaryotes the glutathione-dependent formaldehyde dehydrogenase (GS-FDH or ADH III) is required to control intracellular levels of both S-nitrosoglutathione (GSNO) and S-nitrosothiols (SNOs) (Liu et al., 2001). Moreover, the GS-FDH is conserved from humans to bacteria and the deletion the reductase-encoding gene in mice and yeast abolishes the GSNO-consuming activity and increases susceptibility to a nitrosative challenge.
E. coli harbours the frmRAB gene cluster coding for the glutathione-dependent formaldehyde dehydrogenase and the S-formylglutathione hydrolase that were primary ascribed to formaldehyde detoxification. FrmR is predicted to be a negative regulator of the frmRAB operon. I addition to FrmB, E. coli possesses a second S-formylglutathione hydrolase encoded by yeiG, which in contrast to FrmB is constitutively transcribed (Gonzalez et al., 2006).
Interestingly, the in vitro transcriptome analysis of the re-isolates from the human colonisation study revealed the frmRAB gene cluster to be de-regulated relative to the parent
strain. In two re-isolates (SR12 and KA25) these genes were found among the most strongly up-regulated genes when compared to their ancestor strain. Moreover, this observation was further corroborated on the protein level where FrmA and FrmB could be identified as the most up-regulated proteins in the cytoplasm when compared to strain 83972. These data indicate that the re-isolates SR12 and KA25 might have encountered during patient colonization conditions which demand either formaldehyde or nitric oxide detoxification.
Many extensive studies regarding NO response faced by pathogens have shown a diverse number of genes to be affected (Flatley et al., 2005; Jarboe et al., 2008; Poole et al., 1996;
Pullan et al., 2007). The gene hmpA coding for the NO-inducible flavohaemoglobin was found to be up-regulated in many bacteria during NO detoxification. In our study, we further corroborate these results, and demonstrate that prolonged exposure to host defence factors and most likely elevated concentrations of nitric oxide in the bladder lead to the up-regulation of hmpA expression in E. coli.
The cytoplasmic protein profiles of the re-isolates SR12 and KA25 together with the transcriptome of re-isolate SR12 pinpointed a di-iron protein YtfE to be up-regulated when compared to the parent strain. Interestingly, growth E. coli ytfE mutants is impaired upon nitrosative stress and this effect was even stronger than for hmpA or norV mutants (Justino et al., 2005). NorV is a protein with non-heme di-iron site involved in NO detoxification (Gardner et al., 2002). Recent studies of the same working group revealed that YtfE is involved in iron-sulphur containing protein activity (Justino et al., 2007). Furthermore, YtfE is required during anaerobic respiration under iron-limiting conditions and is hypothesized to be involved in the biosynthesis and repair of iron-sulphur clusters. Thus, the comparison of in vivo re-isolates and the parent strain 83972 revealed another gene, ytfE, which is indirectly involved in stress response and NO detoxification during human bladder colonization.