«(Über die Bedeutung der bakteriellen Genomplastizität für die Adaptation und Evolution asymptomatischer Bakteriurie (ABU) Escherichia coli ...»
Single microfermenters in a form of glass tube with an air and media inflow and a media outflow were placed in a water bath at 37 °C. Fresh media was pumped in by a peristaltic pump from a media tank with the velocity of 500 ml per 24 h. Fresh solutions of nitric oxide containing media (25 mM DETA NOnate, Cayman Chemical, US) were supplied every second day because of the 58 h DETA NOnate half-life. Bacterial colonization of the fresh media reservoir was prevented by placing a hydrophilic filter in front of the each fermenter.
Aeration of the culture was achieved by using an aquarium bubbler. Media and air inflow enhanced media outflow which was collected in an autoclavable tank. All elements were assembled and sterilized by autoclaving. Two days before inoculation, the microfermenter setup was tested for any leakage and potential contamination. Inoculation was performed by injection of 100 µl bacterial overnight culture into each of the microfermenters through a rubber gasket in the fermenter’s screw lid. Sampling was done weekly and bacteria were stored in glycerol stock cultures at –80 °C. At each sampling time point, bacteria were plated on McConkey and Congo Red agar plates in order to access the culture homogeneity.
Moreover, the number of bacteria in the culture was assessed by counting colony forming units on agar plates and by OD600 measurements. Spontaneous mutator phenotype formation was monitored by plating bacteria on a streptomycin agar plates (100 µg/ml).
Fig. 8: Schematic construction of the single flow culture unit. Fresh media was pumped by a peristaltic pump into the microfermenter via hydrophilic filter. Media and air inflow forced media outflow to the waist container.
Fig. 9: Four-chamber microfermenter setup. 1) LB + NO fresh media reservoir; 2) LB fresh media reservoir; 3) Fresh media loading system with a rubber gasket; 4) multi-channel peristaltic pump; 5) single fermenter chamber with a rubber gasket on the top for inoculation; 6) hydrophilic filter to prevent bacterial colonization; 7) fresh media loading reservoir; 8) fresh media loading peristaltic pump; 9) syringe needle for sterile media transfer; 10) sampling system, tubing kept always in ethanol
11) hydrophobic filters for culture aeration system
4.7. In silico analysis For standard sequence comparison and similarity searches, the Basic Local Alignment Search Tool (BLAST) at the National Centre for Biotechnology Information (NCBI) homepage was used. For alignments of nucleotide and amino acid sequences, the BioEdit sequence alignment editor V7.0.1 and VectorNTI V7.0 was used. Genome comparison was performed using the Artemis Comparison Tool (ACT) Release 4 of the Sanger Institute.
5.1. Diversity of clinical ABU E. coli isolates The aim of this part of the study was the comparative geno- and phenotypic analysis of eleven ABU isolates to characterize in detail a larger group of strains of this pathotype and to extend the knowledge on the underlying molecular mechanisms of the ABU lifestyle.
5.1.1. Analysis of relatedness of different ABU isolates Affiliation to the main E. coli phylogenetic lineages revealed that seven of the ABU strains tested belonged to ECOR groups B2 and D which typically include ExPEC isolates. Four isolates, however, represented members of the ECOR groups A and B1 (Table 7). For isolates of the latter phylogenetic groups, it is rather uncommon to be associated with extraintestinal infections.
A more detailed analysis of the phylogenetic relationships by MLST further corroborated the finding that the ABU strains belong to multiple different phylogenetic lineages (Table 7).
Among the strains of the same ECOR group, different non-related clonal groups have been observed. Interestingly, the majority of ABU isolates of ECOR group B2 belongs to the sequence type (ST) 73 which also comprises the well-characterized UPEC O6 isolate CFT073 as well as the non-pathogenic fecal O6 isolate Nissle 1917. The STs 12 and 405 to which also ABU isolates have been allocated, comprise extraintestinal pathogenic as well as nonpathogenic isolates of ECOR group B2 and D, respectively. So far, only non-pathogenic strains belonging to ECOR group B1 have been described for ST 53.
2. Comparative Genomic Hybridization (CGH) In order to compare the genetic diversity of the different ABU E. coli isolates, the genome content was assessed by comparative genomic hybridization using an E. coli K-12 strain MG1655-specific array as well as the E. coli pathoarray. The results obtained from the K-12 array assesses the common genomic content with non-pathogenic strain MG1655, whereas the ”E. coli pathoarray” detects many virulence determinants and island-associated genes present in different ABU genomes.
The CGH results demonstrated a considerable genetic diversity among the eleven ABU isolates tested (Fig. 10; Table 7). On average, 12.9 % of the translatable ORFs present in K-12 strain MG1655 were not detectable in the individual isolates. Based on the functional classification of the GenProtEC database of the chromosomally encoded genes and proteins of E. coli K-12 (http://genprotec.mbl.edu), the majority of these missing ORFs in every strain can be functionally grouped as coding for hypothetical, unclassified or unknown gene products.
Fig. 10: Analysis of the genome content of ABU E. coli isolates. Red and black denote the presence and absence of genes, detected by comparative genomic hybridization respectively. The dendrogram shows the estimated relationships of the different strains obtained by hierarchical cluster analysis of the hybridization signals. The individual STs and major phylogenetic groups of the ABU isolates are indicated.
The eleven E. coli isolates also exhibited a great diversity in ORFs which represent mobile genetic elements or code for structural components of the cell. The alterations were found to be scattered over the entire E. coli MG1655 chromosome. However, prophages of strain MG1655 represent chromosomal variation "hot spots". The individual isolates could be subgrouped according to their CGH barcodes into certain clusters which generally correlate with the main phylogenetic lineage of the individual isolates.
Isolates of the ECOR groups A and B1 harboured markedly less ExPEC-associated genes (on average 9.7 % of the detectable ExPEC genes) than strains that belong to the ECOR groups B2 and D (on average 42.1 % of the detectable ExPEC genes), i.e. those phylogenetic groups that typically comprise extraintestinal pathogenic E. coli. Typical virulence-associated marker genes of intestinal pathogenic E. coli (IPEC) were usually only detected in very low amounts (on average 4.5 % of the detectable IPEC genes). The CGH results of the E. coli pathoarray were partially confirmed by PCR allowing the detection of typical ExPEC-associated and virulence-associated determinants coding for, e.g. different adhesins, toxins, the polyketide colibactin, siderophores, capsules (Table 7). Whereas the fimH gene coding for the adhesin of type 1 fimbriae was detectable in all strains tested, genes of the P- and S/F1C fimbriaeencoding gene clusters are only present in ABU isolates of ECOR group B2. The screening for toxin-, siderophore system- and group II capsule-encoding determinants resulted in similar findings.
These data indicate that ABU isolates differ considerably in their genome content with regard to the presence of virulence-associated genes of uropathogenic E. coli. The virulenceassociated gene content of about two thirds of the ABU strains analyzed resembles that of typical UPEC, whereas in one third of the ABU isolates only a small amount of such determinants exists.
3. Genomic fingerprints of different ABU isolates To further extend the genotypic comparison, the genome structure of the ABU strains was compared by PFGE and rep-PCR. Although the genetic fingerprints of individual isolates belonging to the same ST were very similar (Fig. 11A), the analysis of genomic XbaI restriction fragment patterns by PFGE indicated that members of the same ST differed considerably in their genome content and –structure (Fig. 11B).
Fig. 11: Genomic fingerprints of asymptomatic bacteriuria E. coli isolates. The similarity of the genome structure was assessed by A) Box PCR and B) PFGE following XbaI digestion.
These results demonstrate that ABU isolates do not represent a specialized bacterial clone, but, instead, are a diverse group of strains that evolved independently from different ancestors of different evolutionary E. coli lineages.
5.1.4. Genome size of different ABU isolates The assessment of the genome size by analysis of genomic I-CeuI restriction fragment patterns by PFGE demonstrated that marked genome size differences exist even among strains of the same ST (Fig. 12; Table 8).
Fig. 12: Assessment of the genome size of asymptomatic bacteriuria E. coli isolates by PFGE following I-CeuI digestion.
The genome sizes of ABU isolates of the ECOR groups A and B1 more closely resembled that of non-pathogenic E. coli K-12 strain MG1655 which belongs to ECOR group A as well.
In contrast, those of members of the ECOR group B2 and D were generally larger than that of strain MG1655. However, the genome sizes of ABU isolates belonging to ST73 were, with one exception, always smaller than that of UPEC strain CFT073 which causes symptomatic UTI.
Table 8: Genome size of ABU E. coli isolates
These data demonstrate that ABU isolates differ in genome size and differences exist even among isolates of the same ST. As compared to UPEC strain CFT073, ABU isolates of the same sequence type have a reduced genome size.
5.2. Phenotypic vs. genotypic characteristics of ABU E. coli isolates The carriage of virulence determinants and the ability to express them by different ABU E.
coli isolates was compared. A number of phenotypic tests were performed in combination with polymerase chain reactions where genetic determinants were investigated in more detail.
5.2.1. Type 1 fimbriae
The expression of type 1 fimbriae was tested by yeast agglutination in the presence or absence of mannose. Surprisingly, only four out of eleven tested strains were able to express functional type 1 fimbriae, however, GCH revealed presence of fimH in all of these isolates.
Therefore, the completeness of the fim gene cluster was investigated by PCR-based screening for each individual gene of the fim determinant (fimB to fimH). Accordingly, the complete gene cluster could be detected in strains 5, 20, 57, 62, 63 and 64. The absence of functional type 1 fimbriae in spite of the presence of the complete fim determinant in strains 5 and 57 suggested that these determinants have been inactivated by point mutations. In case of strains 27 and 37, a large 4,253-bp deletion within the fim gene cluster was observed which is also present in strain 83972. Due to this internal deletion including the fimEAICD genes, a truncated fimB gene is fused with a truncated fimD gene probably by recombination between a 7-bp DNA motif GGCGTTT present in both genes. Moreover, in strains 21 and 38, a 29,349bp deletion of large parts of the KpLE2 phage element and the fim operon could be detected.
In these strains, most likely insertion sequence (IS) element-mediated deletion was responsible for the loss of a chromosomal region ranging from a non-functional copy of IS1 upstream of fecI to fimG (Fig. 13). The 29-kb chromosomal region has been replaced by a 1,347-bp DNA stretch which represents a non-functional allele of an IS element, ISEhe3, frequently found, e.g. in Shigella flexneri.
Consequently, the fim determinant of the ABU strains tested represents a heterogenous genomic region which is frequently subjected to point mutations and deletions which cause inactivation of this gene cluster.
Fig. 13: Genetic structure of the fim determinant and adjacent KpLE2 phage-like chromosomal region in asymptomatic bacteriuria E. coli isolates. The scheme is based on the E. coli K-12 chromosome.
Genes of the fim determinant are indicated by hatched arrows, ORFs of the KpLE2 prophage are indicated by filled grey arrows. Dotted arrows represent the fec determinant located within KpLE2.
The ORF A and non-functional ORF B of ISEhe3-like element that replaces large regions of KpLE2 in ABU strains 21 and 38 are indicated by black arrows.
Moreover, analysis of the fimH allelic variation indicated that differences regarding the FimH amino acid sequence were visible between isolates of distinct phylogenetic groups (Fig. 14).
Strains of ECOR groups A and B1 differed from those of groups B2 and D in the presence of valine at position 27 instead of alanine. The ABU isolates that belong to ECOR group B2 exhibit the highest number of amino acid exchanges relative to FimH of E. coli K-12 strain MG1655. Interestingly, marked differences in the FimH amino acid sequence were even observed among the closely related B2 strains of ST73.
Fig. 14: Allelic variation of the FimH type 1 fimbrial adhesins among asymptomatic bacteriuria E. coli isolates. The allocation of the individual strains to the main phylogenetic lineages and clonal groups has been indicated.
5.2.2. P fimbriae To further extend the knowledge about fimbriae expression of the ABU strains, the expression of P fimbriae was tested. PCR screening resulted in five strains positive for the pap fimbrial determinant whereas agglutination with P-fimbriae-specific antibodies was positive only for ABU strain 64. Sequence analysis and comparison of the pap operons of ABU strains 27, 37 63 and 83972 showed that these strains harbour the identical papG allele which codes for a non-functional P fimbrial adhesin.
5.2.3. F1C fimbriae The same approach was used to analyze functionality of F1C fimbriae where agglutination with antibodies specific for F1C fimbriae was positive only for isolates 20, 27 and 64.