In contrast, in case of GI5 we were not able to detect a circular intermediate neither with the originally predicted borders nor with the additional genes suggested by the microarray experiments (Bpet3771–3779), although the microarray data of the phenotypic learn more variants f, g, and k definitely revealed the deletion of this element from their genomes. As shown above, we were

able to detect circular intermediates of most genomic islands by PCR amplification, although the microarray experiments with the phenotypic variants clearly demonstrated the deletion events. Possible explanations for this fact could be that the excised islands are diluted during growth of the bacteria since they cannot replicate. Moreover, the experimental protocols for the two methods are different and PCR amplification is much more sensitive as compared to cy3/cy5 labeling by

Klenow polymerisation. Stability of genomic island GI3 The frequent appearance of phenotypic variants involving the genomic islands AG-120 order present in the B. petrii genome and the detection of circular intermediates of these islands under standard growth conditions indicates that these genomic islands are rather unstable and active at least in terms of excision. To assess the stability of one of these islands (GI3) by homologous recombination we integrated a tetracycline resistance cassette in GI3 between the genes Bpet1523 and Bpet1524 coding for a putative transposase and a glycosyltransferase, respectively. Under standard growth conditions, the resulting strain B. petrii GI3::tetR

did not show any change in its maximum specific growth rate as compared to the wild type (data not shown). This strain was then used for KPT-8602 order growth experiments without selective pressure in which the bacteria were cultivated for about 150 consecutive generations. Exponentially growing B. petrii before has a generation time of about 90 min (data not shown). Figure 5 shows the time course of loss of GI3::tetR determined by differential counting of tetracycline resistant and sensitive bacteria plated out on the respective agar plates. GI3 was stably present in the B. petrii population for about 40 generations, then the proportion of tetracycline resistant bacteria declined steadily and virtually no tetracycline resistant bacteria were found in the population after about 100 generations. Lack of the entire GI3 was confirmed by Southern blotting in representatives of these bacteria (data not shown). Although we cannot exclude a destabilizing effect of the tetracycline cassette on the island, it is likely that GI3 is highly unstable and gets lost with a high incidence when no selective pressure for its persistence is present. Figure 5 Stability of the genomic island GI3 in the genome of B. petrii during culture grown without selective pressure. On the x-axis the number of consecutive generations of the bacteria culture and on the y-axis the proportion of tetracycline resistant bacteria in the culture is shown.