Bacterial RNA synthesis: back to the limelight

Bacteria have been a mainstay of molecular biology, shaping our understanding of the fundamental principles of gene expression control for over half a century. The elegant simplicity of bacterial systems led to many textbook models. Early studies of transcription in bacteria and phages provided a foundation for analysis of more complex eukaryotic systems, and bacterial research started falling out of fashion, with its subjects increasingly seen as over-studied and far removed from modern public-health concerns.

While bacterial systems are indeed simpler – from smaller, more information-packed genomes to fewer subunits in RNA polymerase (RNAP) – part of the simplicity in our explanatory models is due to experimental choices made by those who developed them. Limited by the tools and methods of earlier decades, researchers relied on elementary and direct approaches that nevertheless provided an evergreen source of insights that were generalized across the bacterial kingdom and beyond.

However, bacteria live in complex environments and exchange not only metabolites but also genetic information. Studies of bacteria in exotic niches and extensive communities, from soils to shales to the human gut, prompted the development of new experimental and computational approaches, revealing that bacteria are very diverse, and many “bacterial” stereotypes do not apply to them all. In this Special Focus issue, we present a collection of reviews that reflect the rapidly changing field of bacterial transcription, highlighting the dawning realization that every aspect – the players, their parts, and their purpose in life and evolution – is more complex than we ever imagined.

Key enzymes of the Central Dogma, RNAP and ribosome, are viewed as highly conserved machines. Yet, Miller et al. show that even the best-studied model bacteria, such as Bacillus subtilis and Escherichia coli, have notably diverse RNAPs [1]. Although biochemical and genetic data suggested that they used distinct strategies to regulate RNA synthesis, it took high-resolution cryo-EM structures to make it clear that even their enzymes are different, with two additional auxiliary subunits in B. subtilis “core” RNAP, ε and δ, thought to contribute to the transcription complex stability and disassembly, respectively [1]. Each RNAP has to adapt to the unique needs of its cell, and acquiring additional modules, either as large domain insertions in E. coli or as dissociable subunits, appears to be a common strategy; e.g., bacterial-type chloroplast RNAP apparently needs ten essential subunits to transcribe a ~150-kb genome. New approaches, such as cryo-electron tomography, can capture transcription complexes in their native environments and will no doubt show that bacteria use astonishingly diverse RNAPs and accessory factors.

Coupling of transcription and translation is an accepted paradigm in prokaryotes that lack physical barriers between the two machineries. A model in which RNAP and ribosome are linked by NusG, the only universally conserved transcription factor, supported the ubiquity of the coupling mechanism. However, recent structural and biochemical data, summarized by Webster and Weixlbaumer [2], challenge this view. Coupling can be achieved by different protein bridges; alternatively, simply by synchronizing the rates of transcription and translation, RNAP and ribosome can stay in close proximity, protecting the nascent RNA. While this coupling underpins mRNA quality control in E. coli, it appears totally absent in B. subtilis, where RNAP outruns the ribosome. The discovery that bacteria must completely rewire their regulatory logic when coupling is absent provided researchers with an opportunity to explore its role in diverse phyla by simply examining the architecture of their operons for regulatory signals they may contain. The result suggests that many bacteria do not couple their RNA and protein syntheses, a topic for future studies.

While making the RNA, RNAP also directs its folding. Co-transcriptional folding of the nascent RNA is broadly accepted, and many reports have been published illustrating its importance for proper folding of riboswitches, ribozymes, and other RNA molecules that must adopt complex three-dimensional structures. Until this decade, the focus was on the kinetic control of RNA folding, e.g., by RNAP pausing at strategic positions. Said and Wahl review new evidence in support of a model in which RNAP and transcription factors function as RNA chaperones [3], with recent cryo-EM structures of transcription complexes revealing molecular interactions that control folding of the nascent RNA. Folding can commence in the RNAP exit channel, which accommodates RNA hairpins and other small structures. RNAP-associated proteins can extend the exit channel to promote the formation of RNA structures or constrict the channel to keep the RNA single-stranded, thereby preventing pausing and termination. Said and Wahl show that while very diverse molecular mechanisms can generate similar outcomes, some chaperoning principles are universally conserved.

In addition to ensuring proper RNA folding, RNAP also contributes to the quality control of its DNA template: a transcribing RNAP surveils the DNA for the presence of lesions, which can slow down or stall RNAP. This triggers transcription-coupled repair (TCR), wherein the repair machinery is guided to the site of damage by proteins that interact with the stalled RNAP. Deaconescu describes the detailed mechanism by which the best-studied TCR factor, the double-stranded-DNA translocase Mfd, mediates removal of the stalled RNAP and recruitment of the Uvr machinery [4]. Decades of Mfd studies using genetic, biochemical, genomic, and, more recently, single-molecule biophysics and single-particle cryo-EM analyses, painted a complex picture of highly dynamic interactions among DNA, RNAP, the repair machinery, and Mfd that underpin the role of the latter in TCR. However, accumulating evidence suggests that, even in the absence of DNA damage, Mfd may function as a general transcription regulator and as a pro-mutagenic factor that promotes rapid evolution in different bacterial phyla.

The mechanism of Rho, an archetypal termination factor, is another textbook paradigm that was radically revised last year, as reviewed by Hao et al. [5]. An ATP-powered translocase, Rho can move along RNA, removing any obstacles, such as annealed nucleic-acid strands and proteins, it encounters, even in the absence of RNAP. A conclusion that Rho breaks RNAP interactions with the nascent RNA by force was foregone, but reports that Rho pervasively associates with RNAP during transcription and may inactivate the enzyme did not fit into this simple model. Recent cryo-EM structures of E. coli Rho caught in the act suggest a very different model: Rho, aided by NusA and NusG, hitchhikes on RNAP until a termination signal is encountered, and then induces dramatic conformational changes in RNAP and nucleic acids, eventually dismantling the RNAP active site. Remarkably, this remodeling is driven by Rho interactions with RNAP and Nus factors, whereas the superlative motor activity of Rho is not used at all.

Although bacteria lack nucleosomes, their DNA is highly organized into chromatin-like structures by nucleoid-associated proteins, or NAPs. NAPs are abundant DNA-binding proteins that also protect DNA from uncontrolled expression of xenogenes and from physical damage. Yet, some NAPs moonlight as gene-specific transcription factors, complicating the assignment of a given DNA-binding protein as a NAP. Amemiya et al. summarize the current state of knowledge in this field [6], focusing on the amazing diversity of bacterial NAPs and numerous roles these proteins play in gene expression. While NAPs are in many ways analogous to eukaryotic histones, they actively shape the RNA synthesis program, not merely the chromosome, and are too diverse to be recognized by “sequence gazing.” While emphasizing that a firm line cannot be drawn between proteins that play structural roles and those that play functional ones, Amemiya et al. present a blueprint for identification of potential NAPs in any bacterium for which RNA expression profiles are available [6], to facilitate discovery and characterization of new NAPs.

In bacteria, the nucleoid structure determines gene expression profiles and is defined by NAP-mediated DNA compaction and by torsional stress, which gives rise to supercoiling. NAPs and other proteins that bind DNA can not only directly interfere with RNAP binding to promoters or roadblock the elongating RNAP, but may also exert their effects indirectly, through formation of small DNA loops or large macrodomains. Supercoiling strongly influences protein-–DNA interactions, including those during all steps of transcription, by modulating the DNA duplex structure. In turn, transcription itself generates torsional stress, creating positively and negatively supercoiled topological domains in front of and behind the moving RNAP. Our understanding of the interplay between transcription, roadblocks, and supercoiling has been greatly facilitated by the application of single-molecule techniques, which can control torsional stress on the DNA while monitoring RNA synthesis. Qian et al. [7] summarize these approaches and provide an overview of recent studies of E. coli RNAP elongation against torsion and through physical and topological roadblocks formed by DNA-binding proteins.

Cultivating bacteria in the lab, we tend to forget that in most natural environments, bacteria are just lying in wait for food, not able to divide and grow. While some genes must be expressed to sustain viability, this sporadic activity is not synchronized across the population. Bergkessel describes the challenges that bacteria face under conditions of growth arrest and technical difficulties we face while studying expression of rare RNAs and proteins in very sparse population [8]. She reviews recent progress achieved, despite these challenges and methodological advances that made this progress possible. Changes in expression of transcription factors that bind to RNAP, alarmones, NAPs, etc., that occur during the transition to, and throughout, growth arrest, have been revealed by new methods that enable assessment of very low synthetic activity. As Bergkessel notes, bacteria, including symbionts, likely engage in many processes of keen interest to humans, such as pathogenesis and biodegradation, while barely growing [8]. Thus, understanding how transcription occurs during growth arrest has important practical and medical implications.

This collection of reviews illustrates an amazing diversity of RNA synthesis control mechanisms in bacteria. Recent advancements in this area are due in large part to rapid deployment of new methods, experimental and computational, that not only enable detailed studies of model systems but also open the opportunities to study bacteria living in all places on Earth. We hope that the readers of Transcription will enjoy this progress report on bacterial RNAP and will stay tuned for future exciting discoveries. We extend our gratitude to all the authors for their thought-provoking contributions.