PUB - Publikationen an der Universität Bielefeld

The current methods for characterizing transcriptional regulatory networks are limited to in vivo conditions, providing a snapshot of multiple regulatory interactions simultaneously. To complement these approaches, a new method called ROSE (Run-Off transcription/RNA-sequencing) has been developed for characterizing bacterial promoters genome-wide. This method involves in vitro transcription coupled to transcriptome sequencing specific for native 5′-ends of transcripts. This method is simple and only requires chromosomal DNA, ribonucleotides, RNA polymerase (RNAP) core enzyme, and a specific sigma (σ) factor that recognizes the corresponding promoters that have to be analyzed. ROSE was carried out on Escherichia coli K-12 MG1655 genomic DNA using the E. coli RNAP holoenzyme, which includes σ70. This procedure identified 3,226 transcription start sites, of which 2,167 were also found in in vivo studies and 598 were new. Many new promoters may be repressed under the tested conditions and have not yet been identified in in vivo experiments. Complementary in vivo experiments were conducted using E. coli K-12 strain BW25113 and isogenic transcription factor gene knockout mutants of the transcription factor genes fis, fur, and hns to test this hypothesis. Comparative transcriptome analysis demonstrated that ROSE could identify bona fide promoters repressed in vivo, which resulted in identifying 10, 4, and 9 potential new promoters regulated by Fis, Fur, and H-NS, respectively.

Furthermore, ROSE was utilized to analyze the stringent response mechanism in E. coli K-12 MG1655. The protein DksA and the alarmone ppGpp were added to extend the ROSE method, resulting in 2,205 TSS with differential transcription activity of at least twofold in the presence of DksA+ppGpp. Out of these, 1,387 TSS showed a positive regulation, and 818 TSS showed a negative regulation. The positively regulated genes were primarily involved in stress response mechanisms, such as starvation, pH, or detoxification response, and involved in amino acid biosynthesis, with 15 out of 21 amino acid biosynthesis pathways showing at least one gene regulated by DksA+ppGpp. The genes negatively regulated by DksA+ppGpp were mainly involved in translational processes, including more than 40% of all ribosome assembly or maturation factors and more than 50% of all ribosomal proteins. Furthermore, 83.7% of all tRNA and operons were found to be repressed by at least twofold. Motif analysis of promoter regions either positively or negatively regulated by DksA+ppGpp revealed that negatively regulated promoters had a G+C-rich discriminator motif. In contrast, positively regulated promoters had an A+T-rich discriminator motif.

A comparison of three alarmones, namely pGpp, ppGpp, and pppGpp, has revealed that all three have similar regulative effects in vitro. The alarmones positively affected amino acid biosynthesis genes and starvation, pH, and heat response genes. At the same time, they showed a negative regulation in translational processes, particularly in tRNA regulation. However, a direct comparison showed that ppGpp had the strongest effects of all three. The promoter motif analysis revealed more similarities in promoters activated by DksA+(p)(p)pGpp than in those that were inhibited. Promoters inhibited by DksA+pGpp and DksA+pppGpp also showed more similarities with promoters solely regulated by DksA, especially in their -10 element.

Lastly, ROSE was conducted on the genomic DNA of Corynebacterium glutamicum ATCC 13032 using purified C. glutamicum RNAP core enzyme, which was reconstituted with both σA and σB. A comparison between the RNAP core enzyme and RNAP-σA revealed that only well-conserved promoter motifs were present in the RNAP-σA experiment. This indicated that purified RNAP can be reconstituted with σ factors, which leads to the specific recognition of promoter motifs. ROSE conducted with RNAP-σB enzyme revealed high preferences for a G nucleotide at position -5 in σB-dependent promoters. RNAP-σA mainly recognized genes involved in growth-related processes, while RNAP-σB mainly recognized genes involved in biosynthesis and stress response processes. However, around half of the 817 identified genes for σA and 674 identified genes for σB were recognized by both σ factors, revealing a dual regulation of many genes by σA and σB in C. glutamicum.

This study has demonstrated the wide-ranging applications of the ROSE method, making it a powerful tool for analyzing σ factor and transcription factor regulons in different types of bacteria. This versatility positions ROSE as an ideal bottom-up approach for characterizing transcriptional networks in bacteria, perfectly complementing top-down in vivo transcriptome studies.