Three‐dimensional chromosome re‐modelling: The integral mechanism of transcription regulation in bacteria

Nucleoid‐associated proteins (NAPs) are architectural proteins of the bacterial chromosome and transcription factors that dynamically organise the chromosome and regulate gene expression in response to physicochemical environmental signals. While the architectural and regulatory functions of NAPs have been verified independently, the coupling between these functions in vivo has not been conclusively proven. Here we describe a model NAP – histone‐like nucleoid structuring protein (H‐NS) – as a coupled sensor‐effector that directly regulates gene expression by chromatin re‐modelling in response to physicochemical environmental signals. We outline how H‐NS‐binding partners and post‐translational modifications modulate the role of H‐NS as a transcription factor by influencing its DNA structuring properties. We consolidate our ideas in models of how H‐NS may regulate the expression of the proVWX and hlyCABD operons by chromatin re‐modelling. The interplay between chromosome structure and gene expression may be a common – but, at present, under‐appreciated – concept of transcription regulation in bacteria.


| INTRODUC TI ON
The bacterial chromosome is compacted and organised in the nucleoid of bacterial cells in a manner that allows every gene to be accessible on demand. The organisation is mediated by DNA-binding architectural proteins, commonly referred to as nucleoid-associated proteins, or NAPs, that exert their role by folding, bridging and twisting the chromosome, locally, at the scale of individual operons or regulatory elements, and over longer ranges resulting in compaction of macrodomains and the formation of inter-arm interactions. The chromosome is also compacted by DNA supercoiling that is maintained by DNA topoisomerases and gyrases. Actively transcribed genes and some NAPs behave as supercoil diffusion barriers, restricting supercoiling density to smaller chromosome segments (reviewed in Dame et al. (2020)). The steric effects of chromosome compaction limit the accessibility of open reading frames and regulatory elements, making chromatin structure a direct regulator of transcription. Being the architects of bacterial chromosomes, NAPs inherently play a dual role in the cell, functioning as chromosome structuring proteins and as transcription factors.
The biochemical properties of proteins are affected by fluctuations in temperature, pH and osmolarity. This can result in, among others, the loss or gain of functionality, a change in multimeric state and a change in protein stability. Often, in NAPs, susceptibility to physicochemical changes affects the structure of the protein that, consequently, tunes DNA-binding affinity and the formation of higher order NAP structures. This makes chromosome compaction, organisation and accessibility dependent on the immediate environment of bacterial cells. NAPs, therefore, provide coupled sensoreffector systems that detect changes in the environment as changes to their DNA structuring properties and manifest these as changes in chromosome architecture, gene accessibility and gene expression profiles. This coupling suggests that the complexity of transcription regulation may have evolved around structural remodelling of the chromosome. The interplay between chromosome structure and gene expression is well established in eukaryotes and forms the foundation of the 4D nucleome project (Dekker et al., 2017). In prokaryotic systems, models of the structural regulation of transcription are limited.

| H-NS
The coupled sensor-effector system of Histone-like nucleoid structuring protein (H-NS) is the best studied among bacterial NAPs.
H-NS is a 137 amino acids long NAP that exists as a dimer in solution (Falconi et al., 1988) formed via the interaction between a pair of N-terminal dimerization domains (amino acids 1-41) (Bloch et al., 2003;Esposito et al., 2002;Ueguchi et al., 1996). The dimer is the smallest functional unit of H-NS and preferentially binds AT-rich DNA through its C-terminal DNA-binding domains (amino acids 96-137) (Shindo et al., 1995). This forms a 'nucleation point' from which H-NS dimers multimerise laterally over the DNA via dimer-dimer interaction domains (amino acids 52-84) forming an H-NS:DNA filament ( Figure 1a). The H-NS DNA-binding domains in a filament contact the DNA in the minor groove every 8-17 base pairs at an average of one contact per 10 base pairs (Shen et al., 2022). The contact site is confined to a locus that can be physically accessed by the DNA-binding domain and is distinguished by its narrower minor groove width, lower electrostatic potential and TA step (Shen et al., 2022). H-NS multimerization is co-operative and allows the filament to extend towards lower affinity regions (Arold et al., 2010;Leonard et al., 2009 place a heavy toll on cellular resources, and from horizontally acquired genes that may be toxic (Lamberte et al., 2017;Navarre et al., 2006;Singh et al., 2014). Interestingly, non-specific repression by H-NS at AT-rich regions is key in directing expression from promoters rather than promoter-like sequences. This is because in the competition between H-NS and RNAP for promoter binding, RNAP more efficiently competes H-NS away from AT-rich sites that match the −10 consensus sequence (Singh & Grainger, 2013).
Elongating RNAP, on the other hand, can transcribe across H-NS:DNA by remodelling the structure as a consequence of the force exerted by the actively transcribing polymerase (Dame et al., 2006;Wang et al., 1998). In vivo, remodelling of the H-NS:DNA nucleoprotein relieves silencing of promoters contained within the structure (Rangarajan & Schnetz, 2018). For instance, the insertion of a cassette comprising the constitutive lacUV5 promoter or arabinose-inducible pBAD promoter and the tR1-λN were made for other H-NS repressed promoters (Rangarajan & Schnetz, 2018). While the force exerted by transcribing RNAP is higher than the force required to break H-NS:DNA bridges with optical tweezers (Dame et al., 2006;Wang et al., 1998), in vitro transcription assays on linearised plasmid templates show that DNA:H-NS:DNA bridges, but not H-NS:DNA filaments, still function as effective transcription roadblocks (Kotlajich et al., 2015).
H-NS-mediated DNA bridges promote RNAP pausing and backtracking, increase the dwell times of RNAP in its paused state, and favour Rho-dependent transcription termination by providing a wider kinetic window for Rho function (Kotlajich et al., 2015).

(c) The effect of H-NS binding partners on H-NS:DNA structures
MerR-family regulators that control gene expression by altering promoter structure in response to changes in the concentrations of specific ligands, and CI-type lambda regulators that control gene expression through DNA binding and bridging in response to environmental cues (Brown et al., 2003;Révet et al., 1999).
However, H-NS-mediated regulation is notably more complex, since, in contrast to conventional transcription factors, H-NS does not bind defined sites in defined oligomeric states.
The role of H-NS as a transcription factor is also modulated by Hha (high hemolysin activity) (Baños et al., 2009;Vivero et al., 2008).

| H -NS -LIKEPROTEINS
NAPs that fulfil similar functional roles as H-NS -H-NS-like proteins -have evolved in several bacteria. Lsr2 of Mycobacteria sp., MvaT F I G U R E 2 H-NS may regulate the expression of proU by chromatin remodelling. (a) The regulatory region of the proU operon consists of the upstream regulatory element (URE; orange), 183 bp in length, positioned at −229 to −46 with respect to the σ 70 -dependent proU P2 promoter (black right-angled arrow), and a 217 bp long downstream regulatory element (DRE; orange) at −40 to +177 relative to the P2 promoter. The DRE contains two 10 bp long high-affinity H-NS binding sites (green) at +20 to +30 and +127 to +137 relative to P2. The 1203 bp long proV open reading frame is positioned 60 bp downstream of the P2 promoter. (b) A model of the regulation of proU by the DRE independent of the URE. In the absence of the URE, t-DRE may regulate expression from P2 by interfering with multiple steps in the transcription process. At low osmolarity, the H-NS:DNA nucleoprotein at the DRE, represented here as a filament, may occlude the binding of RNA polymerase (RNAP; blue) to P2, prevent promoter escape, and reduce the processivity of elongating RNAP by trapping positive supercoils in the downstream template (not represented). The increase in intracellular K + at high osmolarity destabilises helix α3 (

(d) A model of the regulation of proU by co-operativity between the URE and DRE
sp. is an atypical H-NS-like protein in that its DNA structuring properties are not affected by physiologically-relevant physicochemical cues in vitro (Erkelens et al., 2022). However, in the presence of small Rok (sRok) as a binding partner, the Rok:DNA nucleoprotein is sensitised to changes in osmolarity (Erkelens et al., 2022) -reminiscent of the interplay between H-NS and StpA (Boudreau et al., 2018).
sRok is a naturally occurring Rok variant that lacks part of the linker between the DNA binding and dimerization domains (Erkelens et al., 2022).

H-NS is a global regulator of gene expression. RNA-Seq studies
show that the deletion of H-NS impacts the relative transcription levels of 5% of genes in the Escherichia coli chromosome (Hommais et al., 2001), while cappable RNA-Seq has shown that in addition to the above, H-NS deletion results in spurious transcripts arising from promoter-like sequences in AT-rich genomic segments (Forrest et al., 2022). An understanding of how transcriptional changes are directly coupled to in vivo chromosome architecture remains elusive.

| TheproVWX(proU)operon
ProVWX (proU) is an H-NS-regulated operon that is activated in response to high osmolarity (Csonka, 1982;Gowrishankar, 1985;Rajkumari et al., 1996). H-NS exerts its effect on proU through the negative regulatory element (NRE) (Dattananda et al., 1991).  (Lucht et al., 1994;Nagarajavel et al., 2007). The DRE, in particular, consists of a pair of highaffinity H-NS-binding sites (5′-TCGATATATT-3′) that closely match the Lang motif for H-NS binding (5′-TCGATAAATT-3′) (Figure 2a) Lang et al., 2007). β-galactosidase expression assays performed by Nagarajavel et al. show that expression from P2 in response to increasing osmolarity in the presence of both the URE and the DRE is sigmoidal. ProU is silenced by the URE and DRE at low osmolarity, and repression is abrogated at higher osmolarities (Nagarajavel et al., 2007). In the presence of only the DRE, expression from P2 increases with osmolarity, but, H-NS exerts its repressive effect at all osmolarities. Interestingly, at low osmolarity, the DRE functions as a weaker repressive element than the URE and DRE combined, but at higher osmolarities, the repression via the DRE only is stronger than with both regulatory elements (Nagarajavel et al., 2007). In the presence of only the URE, H-NS weakly represses P2 at low osmolarity, with a small increase in salt concentration alleviating repression (Nagarajavel et al., 2007). The repressive effects on proU P2 mediated by H-NS via the URE and the DRE are not additive, suggesting that they function cooperatively.
Cooperativity may be observed if the URE and DRE separately repress different steps of transcription -initiation and elongation, respectively. However, cooperativity between a pair of regulatory elements that are bound by an architectural protein -H-NS -may also be achieved by a physical interaction.
Mechanistically, the DRE may repress P2 by interfering with both, transcription initiation and elongation. In the absence of the URE, the H-NS:DNA filament formed at the DRE may occlude the binding of RNAP to the promoter and reduce the processivity of RNAP by behaving as a roadblock (Figure 2b). Increasing extracellular osmolarity triggers the influx of K + (Sleator & Hill, 2002), and up-regulates expression from P2 due to the inherent osmosensitivity of the pro-

| ThehlyCABD(hly)operon
The hlyCABD operon is encoded by uropathogenic E. coli strains that express the hemolysin toxin (Goebel & Schrempf, 1971). HlyCABD is osmolarity-and temperature-sensitive and is co-regulated by H-NS and Hha (Mouriño et al., 1994). H-NS binds the regulatory region of hlyCABD, which is positioned at the 5′ end of the operon extending into hlyC, and represses hly. Hha reinforces H-NS mediated repression. In vivo transcription assays show that Hha can repress hly even in a Δhns background (Nieto et al., 2000). Owing to the absence of a DNA-binding domain in Hha, and because Hha mediates its effect via H-NS (Ali et al., 2013), in Δhns strains, Hha may function via the H-NS paralogue -and molecular back-up -StpA.
In vivo, hlyCABD is repressed by high osmolarity (Nieto et al., 2000). The repression is weakly alleviated by the deletion of hha and strongly alleviated in Δhns strains. Expression is highest in a Δhns Δhha background (Nieto et al., 2000). At high osmolarity, helix  (Nieto et al., 2000). The presence of StpA in H-NS:DNA structures in vivo (Uyar et al., 2009) may contribute to the stability of the nucleoprotein complex (Boudreau et al., 2018). Hly is also repressed by high osmolarity in a Δhns Δhha strain (Nieto et al., 2000), indicating that either an additional hly repressor that exerts its effect at high osmolarity exists, or the hly promoter is inherently osmosensitive and is repressed by high osmolarity -reminiscent of the proVWX P2 promoter that is, conversely, activated by high osmolarity in a purified in vitro transcription system lacking additional transcription factors (Rajkumari et al., 1996). Due to the complexity of an in vivo system, the precise roles of H-NS and Hha and the potential roles of StpA and an osmosensitive promoter are difficult to discern. Disentangling the effects of these regulators on hly will require in vitro transcription studies.
H-NS mediates the repression regardless of the genetic context of Hha (Nieto et al., 2000). At lower temperatures, the multimeric, nonsequestered structure of H-NS is more prevalent due to the stabili-

| THEELUS IVECHROMOSOME
The models presented above to describe the interplay between local three-dimensional chromosome structure and transcription are proposed based on in vivo transcription assays (Nagarajavel et al., 2007;Nieto et al., 2000), in vitro biochemical and biophysical techniques (Ali et al., 2013;Arold et al., 2010;Boudreau et al., 2018;Hameed et al., 2019;Kotlajich et al., 2015;van der Valk et al., 2017), and in Investigating the re-modelling of bacterial operons requires an assay that provides a resolution of individual regulatory regions that may be shorter than 100 bp. At the present time, this is difficult and costly to achieve with high-throughput assays such as 3C-Seq and Hi-C.
However, the resolution can be achieved with 3C-qPCR. 3C-qPCR is a low-throughput assay that probes the relative interaction frequency between a pair of loci at the resolution of individual restriction fragments, and therefore, offers a reliable system to probe local chromosome structure (Hagege et al., 2007;Rashid et al., 2022). 3C-qPCR can conclusively prove if the local structure of the chromosome is remodelled in response to changes in the environment, and, in association with RT-qPCR can verify if these structural changes involve a change in transcription.

| CON CLUS I ONANDPER S PEC TIVE S
Bacteria have evolved to respond to physicochemical environmental cues selectively on a transcriptional level. NAPs are key in this response. The dependence of the biochemical and architectural properties of NAPs on the milieu tether the physical environment to a biological system. The consequential re-modelling of the chromosome in response to the environment and the steric changes in gene accessibility manifest as altered transcription profiles. This highlights that NAPs directly link a homeostatic system to the fickle environment to which it needs to respond to survive. The diversity of NAPs that respond to different physicochemical cues, and the evolution of mechanisms that modulate the role of NAPs by fine-tuning their architectural properties (binding partners and post-translational modifications), suggests that the principle of chromatin remodelling to regulate gene expression in bacteria may be ubiquitous. Towards a complete understanding of transcription regulation in bacteria, gene expression cannot be uncoupled from chromatin structure.

ACK N O WLE D G E M ENTS
Research in the lab of RTD is supported by grants from the Netherlands Organisation for Scientific Research (VICI 016.160.613/533 and OCENW.GROOT.2019.012) (RTD).

CO N FLI C TO FI NTE R E S TS TATE M E NT
The authors declare no conflicts of interest.

DATAAVA I L A B I L I T YS TAT E M E N T
Data sharing is not applicable to this article as no new data were created or analyzed in this study.