MYC is overexpressed in most types of tumors, but a means to selectively decrease its expression is yet to be found. Our recent findings on modulation of BCL2 gene expression through protein interactions with the BCL2 i-motif have provided a basis for further investigation of MYC gene control. It is proposed that the MYC i-motif could function by a similar molecular switch mechanism as in BCL2.Binding sites for heterogeneous nuclear ribonucleoprotein K (hnRNP K) within the MYC promoter also exist in the i-motif-forming sequence. Circular dichroism and bromine footprinting confirmed that this DNA sequence is able to form an i-motif, and systematic mutation of the cytosine residues in this sequence has revealed a 5:5:5 loop configuration. Indeed, all loops of the i-motif, when folded into a 5:5:5 loop configuration, contain the hnRNP K consensus sequence (CCCT). Previous studies show that hnRNP K binds to this i-motif-forming sequence, but it was assumed to be single-stranded. Binding studies revealed that hnRNP K has more binding affinity to its consensus sequence in the i-motif compared to a mutant sequence where the i-motif cannot form. Further investigation of the MYC promoter revealed an additional two runs of cytosine seven bases downstream of the MYC i-motif. Biophysical studies showed that the additional two runs were not involved in i-motif formation, however recent studies describe their importance for transcriptional activation. We found that hnRNP K preferred the longer 5CT sequence compared to the i-motif forming 4CT sequence when using a competitive binding assay. Utilizing luciferase reporters containing either the 4CT or 5CT sequence validated that hnRNP K required both the i-motif and 5th CT element for maximum transcriptional activation. Competition binding studies and bromine footprinting showed that hnRNP K bound to the downstream 5th CT element and the central and lateral loops of the i-motif.Additionally, we found that co-overexpression of Sp1 and hnRNP K induced a 10-fold increase in luciferase activity in the 5CT reporter only. We hypothesize that Sp1 continuously primes the promoter to initiate transcription inducing more negative superhelicity and increasing the melting of duplex DNA. This increased melting grants hnRNP K’s three KH domains access to the i-motif loops and the 5Th CT element. Confirmation by ChIP analysis validated that Sp1 overexpression causes an increase in hnRNP K occupancy at the MYC promoter. These findings provide new insight into the mechanisms of MYC transcriptional control by the i-motif and G-quadruplex.Recently, our group has demonstrated that two small molecules IMC-48 and IMC-76 can interact with the i-motif and can be an effective means to modulate BCL2 expression. Based on these results with the BCL2 i-motif, we employed a similar strategy and screened and identified small drug-like molecules that interact with MYC i-motif, using a FRET high-throughput assay. We then further validated that IMC-16 stabilizes the MYC i-motif through the interactions with the loops of the i-motif. No stabilization by IMC-16 treatment was observed with the MYC G-quadruplex and the BCL2 and PDGFRβi-motifs demonstrating selectivity for the MYC i-motif.Finally, we investigated the effects of IMC-16 on MYC expression in three lymphoma cell lines all expressing different levels of MYC. In the case of both Daudi and RAJI Burkitt’s lymphoma cell lines we demonstrated that selectively stabilizing the i-motif by IMC-16 could increase MYC expression. Furthermore, we demonstrated that the MYC G-quadruplex stabilizing compound GQC-05 and IMC-16, which stabilizes the MYC i-motif, have antagonistic effects on MYC expression, providing further evidence of a molecular switch mechanism in the NHEIII1. Directly targeting MYC expression through the i-motif offers advantages over targeting the G-quadruplex, because of the reduced stability and dynamic nature of the i-motif, additionally the i-motif is only found in DNA. The use of such i-motif interactive compounds is the first step into the development of new innovative approaches to treat cancers.

Within every human cell, approximately two meters of DNA must be compacted into a nucleus with a diameter of around ten micrometers. Alongside this daunting storage problem, the 3D organisation of the genome also helps determine which genes are up- or down-regulated, which in turn effects the functionality of the cell itself. While the organisational structure of the genome can be revealed using experimental techniques such as chromosome conformation capture and its high-throughput variant Hi-C, the mechanisms driving this organisation are still unclear. The first two results chapters of this thesis use molecular dynamics simulations to investigate the effect of a potential organisational mechanisms for DNA known as the "bridging-induced attraction". This mechanism involves multivalent DNA-binding proteins bridging genomically distant regions of DNA, which in turn promotes further binding of proteins and compaction of the DNA. In chapter 2 (the first results chapter) we look at a model where proteins can bind non-specifically to DNA, leading to cluster formation for suitable protein-DNA interaction strengths. We also show the effects of protein concentration on the DNA, with a collapse from a swollen to a globular phase observed for suitably high protein concentrations. Chapter 3 develops this model further, using genomic data from the ENCODE project to simulate the "specific binding" of proteins to either active (euchromatin) or inactive (heterochromatin) regions. We were then able to compare contact maps for specific simulated chromosomes with the experimental Hi-C data, with our model reproducing well the topologically associated domains (TADs) seen in Hi-C contact maps. In chapter 4 of the thesis we use numerical methods to study a model for the coupling between DNA topology (in particular, supercoiling in DNA and chromatin) and transcription in a genome. We present details of this model, where supercoiling flux is induced by gene transcription, and can diffuse along the DNA. The probability of transcription is also related to supercoiling, as regions of DNA which are negatively supercoiled have a greater likelihood of being transcribed. By changing the magnitude of supercoiling flux, we see a transition between a regime where transcription is random and a regime where transcription is highly correlated. We also find that divergent gene pairs show increased transcriptional activity, along with transcriptional waves and bursts in the highly correlated regime { all these features are associated with genomes of living organisms.

Escherichia coli est capable de survivre dans de nombreux environnements différents. Les informations nécessaires à cette adaptation sont codées dans le chromosome. Cette molécule circulaire est condensé dans une structure compacte protéines-ADN, appelée nucléoïde. Le chromosome n¿est pas uniforme et montre notamment une distribution inégale de sites de fixation de protéines et de séquences riches en AT. Il a été montré que la position des gènes importants pour la cellule est hautement conservée dans les gamma-protéobactéries. Ces différences le long du chromosome et cette conservation de la position suggèrent que la position du gène peut influencer son expression. Pour tester cette hypothèse, on a étudié l'expression d'un gène fluorescent inséré dans différentes positions autour du chromosome. L'expression de ce gène est contrôlé par des promoteurs différemment régulés: un est réprimé par la protéine H-NS, un est non régulé et un est sensible au superenroulement de l'ADN. Nous avons étudié l'expression dynamique de ces promoteurs pendant les différentes phases de croissance dans différentes conditions. Nous avons montré que l'expression du promoteur dépendant de la protéine H-NS est liée à l'emplacement sur le chromosome. En effet, la répression par H-NS est accrue en présence de séquences riches en AT. Nous avons également étudié l'influence d'un gène divergent sur l'expression de gènes rapporteurs en fonction de la position chromosomique. Nous avons montré que cette influence dépend de la localisation du gène. Nous avons donc demontré l'impact de la position chromosomique sur l'expression des gènes tout en donnant une nouvelle perspective sur la complexité du génome. / Escherichia coli is able to survive in many different environments. The information necessary for this adaptation is encoded in the chromosome. This circular molecule is condensed in a compact DNA-protein structure, called the nucleoid. The chromosome is not uniform, and shows uneven distributions of nucleoid-associated proteins (NAPs) binding sites, AT-rich sequences and general protein occupancy domains. It has been demonstrated that the position of important genes is highly conserved in ?-Proteobacteria. These differences along the chromosome and the conserved position of important genes suggest that the position of the gene can influence gene expression. To test this hypothesis, I studied the expression of a fluorescent reporter gene inserted in different positions around the chromosome. The expression of the reporter is driven by differently regulated promoters, one repressed by the important NAP H-NS, one non regulated and one subject to supercoiling and stringent control. We studied the dynamical expression of these promoters in different growth conditions, growth phases, upon nutritional upshift and under stress. We showed that the expression of the H-NS dependent promoter depends on the location on the chromosome, because H-NS repression is enhanced in presence of AT-rich sequences. We also studied the influence of a divergent gene on the reporter expression as a function of chromosomal position, and showed that this influence depends on the location of the gene. With our study we have been therefore able to show the impact of chromosomal position on gene expression and to give a new perspective on genome complexity.

No description available.

Des variations importantes du surenroulement de l’ADN peuvent être générées durant la phase d’élongation de la transcription selon le modèle du « twin supercoiled domain ». Selon ce modèle, le déplacement du complexe de transcription génère du surenroulement positif à l’avant, et du surenroulement négatif à l’arrière de l’ARN polymérase. Le rôle essentiel de la topoisomérase I chez Escherichia coli est de prévenir l’accumulation de ce surenroulement négatif générée durant la transcription. En absence de topoisomérase I, l’accumulation de ce surenroulement négatif favorise la formation de R-loops qui ont pour conséquence d’inhiber la croissance bactérienne. Les R-loops sont des hybrides ARN-ADN qui se forment entre l’ARN nouvellement synthétisé et le simple brin d’ADN complémentaire. Dans les cellules déficientes en topoisomérase I, des mutations compensatoires s’accumulent dans les gènes qui codent pour la gyrase, réduisant le niveau de surenroulement négatif du chromosome et favorisant la croissance. Une des ces mutations est une gyrase thermosensible qui s’exprime à 37 °C. La RNase HI, une enzyme qui dégrade la partie ARN d’un R-loop, peut aussi restaurer la croissance en absence de topoisomérase I lorsqu’elle est produite en très grande quantité par rapport à sa concentration physiologique. En présence de topoisomérase I, des R-loops peuvent aussi se former lorsque la RNase HI est inactive. Dans ces souches mutantes, les R-loops induisent la réponse SOS et la réplication constitutive de l’ADN (cSDR). Dans notre étude, nous montrons comment les R-loops formés en absence de topoisomérase I ou RNase HI peuvent affecter négativement la croissance des cellules. Lorsque la topoisomérase I est inactivée, l’accumulation d’hypersurenroulement négatif conduit à la formation de nombreux R-loops, ce qui déclenche la dégradation de l’ARN synthétisé. Issus de la dégradation de l’ARNm de pleine longueur, des ARNm incomplets et traductibles s’accumulent et causent l’inhibition de la synthèse protéique et de la croissance. Le processus par lequel l’ARN est dégradé n’est pas encore complètement élucidé, mais nos résultats soutiennent fortement que la RNase HI présente en concentration physiologique est responsable de ce phénotype. Chose importante, la RNase E qui est l’endoribonuclease majeure de la cellule n’est pas impliquée dans ce processus, et la dégradation de l’ARN survient avant son action. Nous montrons aussi qu’une corrélation parfaite existe entre la concentration de RNase HI, l’accumulation d’hypersurenroulement négatif et l’inhibition de la croissance bactérienne. Lorsque la RNase HI est en excès, l’accumulation de surenroulement négatif est inhibée et la croissance n’est pas affectée. L’inverse se produit Lorsque la RNase HI est en concentration physiologique. En limitant l’accumulation d’hypersurenroulement négatif, la surproduction de la RNase HI prévient alors la dégradation de l’ARN et permet la croissance. Quand la RNase HI est inactivée en présence de topoisomérase I, les R-loops réduisent le niveau d’expression de nombreux gènes, incluant des gènes de résistance aux stress comme rpoH et grpE. Cette inhibition de l’expression génique n’est pas accompagnée de la dégradation de l’ARN contrairement à ce qui se produit en absence de topoisomérase I. Dans le mutant déficient en RNase HI, la diminution de l’expression génique réduit la concentration cellulaire de différentes protéines, ce qui altère négativement le taux de croissance et affecte dramatiquement la survie des cellules exposées aux stress de hautes températures et oxydatifs. Une inactivation de RecA, le facteur essentiel qui déclenche la réponse SOS et le cSDR, ne restaure pas l’expression génique. Ceci démontre que la réponse SOS et le cSDR ne sont pas impliqués dans l’inhibition de l’expression génique en absence de RNase HI. La croissance bactérienne qui est inhibée en absence de topoisomérase I, reprend lorsque l’excès de surenroulement négatif est éliminé. En absence de RNase HI et de topoisomérase I, le surenroulement négatif est très relaxé. Il semble que la réponse cellulaire suite à la formation de R-loops, soit la relaxation du surenroulement négatif. Selon le même principe, des mutations compensatoires dans la gyrase apparaissent en absence de topoisomérase I et réduisent l’accumulation de surenroulement négatif. Ceci supporte fortement l’idée que le surenroulement négatif joue un rôle primordial dans la formation de R-loop. La régulation du surenroulement négatif de l’ADN est donc une tâche essentielle pour la cellule. Elle favorise notamment l’expression génique optimale durant la croissance et l’exposition aux stress, en limitant la formation de R-loops. La topoisomérase I et la RNase HI jouent un rôle important et complémentaire dans ce processus. / Important fluctuations of DNA supercoiling occur during transcription in the frame of the “twin supercoiled domain” model. In this model, transcription elongation generates negative and positive supercoiling respectively, upstream and downstream of the moving RNA polymerase. The major role of bacterial topoisomerase I is to prevent the accumulation of transcription-induced negative supercoiling. In its absence, the accumulation of negative supercoiling triggers R-loop formation which inhibits bacterial growth. R-loops are DNA/RNA hybrids formed during transcription when the nascent RNA hybridizes with the template strand thus, leaving the non-template strand single stranded. In cells lacking DNA topoisomerase I, a constant and selective pressure for the acquisition of compensatory mutations in gyrase genes reduces the negative supercoiling level of the chromosome and allows growth. One of these mutations is a thermosensitive gyrase expressed at 37 °C. The overexpression of RNase HI, an enzyme that degrades the RNA moiety of an R-loop, is also able to correct growth inhibition in absence of topoisomerase I. In the presence of topoisomerase I, R-loops can also form when RNase HI is lacking. In these mutants, R-loop formation induces SOS and constitutive stable DNA replication (cSDR). In our study, we show how R-loops formed in cells lacking topoisomerase I or RNase HI can affect bacterial growth. When topoisomerase I is inactivated, the accumulation of hypernegative supercoiling inhibits growth by causing extensive R-loop formation which, in turn, can lead to RNA degradation. As a result of RNA degradation, the accumulation of truncated and functional mRNA instead of full length ones, is responsible for protein synthesis inhibition that alters bacterial growth. The mechanism by which RNA is degraded is not completely clear but our results strongly suggest that RNase HI is involved in this process. More importantly, the major endoribonuclease, RNase E, is not involved in RNA degradation because RNA is degraded before its action. We show also that there is a perfect correlation between RNase HI concentration, the accumulation of hypernegative supercoiling and bacterial growth inhibition. When RNase HI is in excess, no accumulation of hypernegative supercoiling and growth inhibition are observed. The opposite is true when RNase HI is at its wild type level. By preventing the accumulation of hypernegative supercoiling, the overproduction of RNase HI inhibits extensive R-loop formation and RNA degradation, thus, allowing growth. In absence of RNase HI (rnhA) and in presence of topoisomerase I, R-loops are also responsible for an inhibition in gene expression, including stress genes such as rpoH and grpE. The inhibition of gene expression is not related to RNA degradation as seen in absence of topoisomerase I but it is rather related to a reduction in gene expression. In absence of RNase HI, the diminution of genes expression is responsible for a reduction in the cellular level of proteins, which negatively affects bacterial growth and bacterial survival to heat shock and oxydative stress. Additional mutations in RecA, the protein that activates SOS and cSDR after R-loop formation in rnhA, do not correct this phenotype in rnhA. Thus, SOS and cSDR are not directly involved in the inhibition of gene expression in the absence of RNase HI. In absence of topoisomerase I, growth inhibition resumes when hypernegative supercoiling is reduced. When compared to wild type strains, DNA is very relaxed in absence of RNase HI and topoisomerase I. It seems that R-loop formation induces the relaxation of negatively supercoiled DNA. All this strongly supports the idea that negative supercoiling plays an important role in R-loop formation. Finally, our work shows how essential negative supercoiling regulation is for cell physiology. By preventing R-loop formation, regulation of negative supercoiling allows optimal gene expression, which is crucial for cellular growth and for stress survival. Both topoisomerase I and RNase HI play an important and complementary role in this process.

No description available.

Computer simulations provide a potentially powerful complement to conventional experimental techniques in elucidating the structures, dynamics and interactions of macromolecules. In this thesis, I present three applications of computer simulations to investigate important biomolecules with sizes ranging from two-residue peptides, to proteins, and to whole chromosome structures. First, I describe the results of 441 independent explicit-solvent molecular dynamics (MD) simulations of all possible two-residue peptides that contain the 20 standard amino acids with neutral and protonated histidine. 3JHNHα coupling constants and δHα chemical shifts calculated from the MD simulations correlated quite well with recently published experimental measurements for a corresponding set of two-residue peptides. Neighboring residue effects (NREs) on the average 3JHNHα and δHα values of adjacent residues were also reasonably well reproduced. The intrinsic conformational preferences of each residue, and their NREs on the conformational preferences of adjacent residues, were analyzed. Finally, these NREs were compared with corresponding effects observed in a coil library and the average β-turn preferences of all residue types were determined. Second, I compare the abilities of three derivatives of the Amber ff99SB force field to reproduce a recent report of 3JHNHα scalar coupling constants for hundreds of two-residue peptides. All-atom MD simulations of 256 two-residue peptides were performed and the results showed that a recently-developed force field (RSFF2) produced a dramatic improvement in the agreement with experimental 3JHNHα coupling constants. I further show that RSFF2 also improved modestly agreement with experimental 3JHNHα coupling constants of five model proteins. However, an analysis of NREs on the 3JHNHα coupling constants of the two-residue peptides indicated little difference between the force fields’ abilities to reproduce experimental NREs. I speculate that this might indicate limitations in the force fields’ descriptions of nonbonded interactions between adjacent side chains or with terminal capping groups. Finally, coarse-grained (CG) models and multi-scale modeling methods are used to develop structural models of entire E. coli chromosomes confined within the experimentally-determined volume of the nucleoid. The final resolution of the chromosome structures built here was one-nucleotide-per-bead (1 NTB), which represents a significant increase in resolution relative to previously published CG chromosome models, in which one bead corresponds to hundreds or even thousands of basepairs. Based on the high-resolution final 1 NTB structures, important physical properties such as major and minor groove widths, distributions of local DNA bending angles, and topological parameters (Linking Number (Lk), Twist (Tw) and Writhe (Wr)) were accurately computed and compared with experimental measurements or predictions from a worm-like chain (WLC) model. All these analyses indicated that the chromosome models built in this study are reasonable at a microscopic level. This chromosome model provides a significant step toward the goal of building a whole-cell model of a bacterial cell.

Transcription by RNA polymerase can induce the formation of hypernegatively supercoiled DNA both in vivo and in vitro. This phenomenon has been explained by a “twin-supercoiled-domain” model of transcription where a positively supercoiled domain is generated ahead of the RNA polymerase and a negatively supercoiled domain behind it. In E. coli cells, transcription-induced topological change of chromosomal DNA is expected to actively remodel chromosomal structure and greatly influence DNA transactions such as transcription, DNA replication, and recombination. In this study, an IPTG-inducible, two-plasmid system was established to study transcription-coupled DNA supercoiling (TCDS) in E. coli topA strains. By performing topology assays, biological studies, and RT-PCR experiments, TCDS in E. coli topA strains was found to be dependent on promoter strength. Expression of a membrane-insertion protein was not needed for strong promoters, although co-transcriptional synthesis of a polypeptide may be required. More importantly, it was demonstrated that the expression of a membrane-insertion tet gene was not sufficient for the production of hypernegatively supercoiled DNA. These phenomenon can be explained by the “twin-supercoiled-domain” model of transcription where the friction force applied to E. coli RNA polymerase plays a critical role in the generation of hypernegatively supercoiled DNA. Additionally, in order to explore whether TCDS is able to greatly influence a coupled DNA transaction, such as activating a divergently-coupled promoter, an in vivo system was set up to study TCDS and its effects on the supercoiling-sensitive leu-500 promoter. The leu-500 mutation is a single A-to-G point mutation in the -10 region of the promoter controlling the leu operon, and the AT to GC mutation is expected to increase the energy barrier for the formation of a functional transcription open complex. Using luciferase assays and RT-PCR experiments, it was demonstrated that transient TCDS, “confined” within promoter regions, is responsible for activation of the coupled transcription initiation of the leu-500 promoter. Taken together, these results demonstrate that transcription is a major chromosomal remodeling force in E. coli cells.

L’entérobactérie Dickeya dadantii est responsable de la maladie de la pourriture molle sur de nombreux hôtes végétaux. Ce symptôme est essentiellement dû à la production d’un arsenal d’enzymes qui dégradent la pectine, ciment des parois des cellules végétales. Parmi ces enzymes, les pectate lyases (Pels) ont un rôle majeur dans le pouvoir pathogène en raison de leur capacité àreproduire, sous forme purifiée, le symptôme de la pourriture molle. La synthèse des Pels est soumise à un contrôle très fin qui fait intervenir différents régulateurs agissant de manière intégrée via un réseau de régulation. De nombreuses conditions environnementales modulent la synthèse des Pels via l’action de ces régulateurs. La température est un facteur qui agit sur leur synthèse et pour lequel les mécanismes moléculaires restaient non élucidés. Lors de cette étude, nous avons montré que le régulateur PecT, un répresseur du réseau de régulation, intervient dans la thermorégulation de la synthèse des Pels. PecT s’est avéré être également impliqué dans la thermorégulation de deux autres fonctions de virulence : la mobilité et la synthèse desexopolysaccharides de surface. La quantification des transcrits des gènes de ces 3 fonctions de virulence a permis de montrer que l’action de PecT dans ce contrôle a lieu au niveau transcriptionnel. Le mécanisme moléculaire de la thermorégulation exercée par PecT a été étudié plus en détail sur les gènes pel. Des résultats obtenus in vivo ont montré que la fixation de PecT sur les régionsrégulatrices des gènes pel est plus efficace quand la température augmente. La croissance de D. dadantii à hautes températures induit un relâchement de l’ADN. De manière remarquable, un relâchement artificiel de l’ADN par un traitement inhibant la gyrase entraine une augmentation de la fixation de PecT sur les gènes pel même pour des cellules cultivées à basses températures. De plus, la délétion de PecT se traduit par une augmentation de la capacité de D. dadantii à induire la pourriture molle à hautes températures. Ainsi la topologie de l’ADN et PecT agissent de manière concertée pour moduler la synthèse des Pels en fonction de la température.L’ensemble de ces données apporte une preuve supplémentaire de l’importance de la dynamique structurale de la chromatine dans l’ajustement de la physiologie bactérienne en réponse aux variations des conditions environnementales. / Bacteria are colonizers of various environments and host organisms, and they are often subjected to drastic temperature variations. Dickeya dadantii is a Gram-negative pathogen infecting a wide range of plant species. Soft rot, the visible symptom, is mainly due to the production of pectate lyases (Pels) that can destroy the plant cell walls. Production of Pels is controlled by a complex regulation system and responds tovarious stimuli, such as presence of pectin, plant extracts, growth phase, temperature or iron concentration. Although many studies have been carried out, the mechanisms of control of Pels production by temperature have not yet been elucidated. In bacteria, thermoregulation acting at the level of transcription initiation occurs usually both via transcription factors and DNA topology. We show that PecT, a previously identified repressor, is involved in the thermoregulation of the pel gene expression. Using in vivo Chromatin ImmunoPrecipitation (ChIP) coupled to quantitative RT-PCR(qRT-PCR), we reveal that PecT binding to the pel gene promoters is modulated by temperature. By manipulating the DNA topology in vivo, we further show that DNA supercoiling state is involved in the thermoregulation of pel gene expression by PecT. In addition, we show that the development of the pathogenicity of the pecT mutant according to changes in temperature is different from that of the parental strain. This report presents a new example of how plant pathogenic bacteria use transcription factor and DNA topology to adjust synthesis of virulence factors in response to temperature variation.

In this thesis, the response of DNA and RNA to linear and torsional mechanical stress is studied using coarse-grained models. Inspired by single-molecule assays developed over the last two decades, the end-to-end extension, buckling and torque response behaviour of the stressed molecules is probed under conditions similar to experimentally used setups. Direct comparison with experimental data yields excellent agreement for many conditions. Results from coarse-grained simulations are also compared to the predictions of continuum models of linear polymer elasticity. A state diagram for supercoiled DNA as a function of twist and tension is determined. A novel confomational state of mechanically stressed DNA is proposed, consisting of a plectonemic structure with a denaturation bubble localized in its end-loop. The interconversion between this novel state and other, known structural motifs of supercoiled DNA is studied in detail. In particular, the influence of sequence properties on the novel state is investigated. Several possible implications for supercoiled DNA structures in vivo are discussed. Furthermore, the dynamical consequences of coupled denaturation and writhing are studied, and used to explain observations from recent single molecule experiments of DNA strand dynamics. Finally, the denaturation behaviour, topology and dynamics of short DNA minicircles is studies using coarse-grained simulations. Long-range interactions in the denaturation behaviour of the system are observed. These are induced by the topology of the system, and are consistent with results from recent molecular imaging studies. The results from coarse-grained simulations are related to modelling of the same system in all-atom simulations and a local denaturation model of DNA, yielding insight into the applicability of these different modelling approaches to study different processes in nucleic acids.